Bidirectional Innovation Spillovers from Focal Firms to Upstream and Downstream Supply-Chain Firms: Evidence from China

Abstract

This paper uses the supply-chain portfolio screened from the key industrial chains and focal firms announced by Chinese governments from 2009 to 2024 as the research sample to examine the bidirectional innovation spillovers from focal firms to supply-chain firms. The research results show that there are bidirectional innovation spillovers from focal firms to supply-chain firms. And the innovation spillover effect is asymmetric; that is, compared with the upstream supply chain, the innovation spillover from focal firms to supply-chain firms in the downstream supply chain is stronger. The mechanism test results show that focal firms can enhance the innovation level of upstream firms through the cooperative innovation mechanism, and can also enhance the innovation level of downstream firms through the knowledge spillover mechanism. This study not only enriches the existing literature on the spillover effects of innovation and the role of focal firms in the supply chain, but also provides empirical evidence and theoretical support for focal firms to play a core role, promote innovation in the supply chain and lead the overall development of the supply chain.

1. Introduction

Innovation, as the core driving force for the survival and development of enterprises, is crucial for enhancing product competitiveness, adapting to rapid market changes, and responding to the evolving demands of consumers [1]. The view that innovation has a spillover effect has become a consensus in the academic community. Innovation can promote technological diffusion and efficiency improvement through various forms and channels such as geographical space, interaction among entities, and legal policies [2,3,4]. Among them, the supply chain has become an important channel for knowledge transfer and technological diffusion [5]. For instance, Toyota established a joint research center with its component suppliers to jointly develop power technology. Taiwan Semiconductor Manufacturing Company (TSMC) provided manufacturing technologies and IP resources to downstream chip design companies through an open innovation platform (OIP), reducing their innovation barriers. Pfizer shared vaccine research data with upstream raw material suppliers and downstream hospitals through a digital collaboration platform, achieving collaborative innovation. Therefore, examining the spillover effect of innovation based on supply chain relationships holds significant theoretical and practical significance.

The focal firms in the supply chain are the core nodes, occupying the central position in terms of technology, funds, and markets [6]. Based on the focal firms, the supply chain forms a division of labor and a collaboration relationship among upstream firms–focal firms–downstream firms. Among them, upstream firms and focal firms constitute the upstream supply chain, while focal firms and downstream firms constitute the downstream supply chain. The existing literature suggests that focal firms in the supply chain can integrate internal and external resources, connect upstream and downstream firms, and generate strong control and influence, leading the overall development of the supply chain [7]. Then, in the supply chain relationship, do focal firms have the bidirectional innovation spillovers to upstream and downstream supply-chain firms? Further, the focal firms occupy different ecological positions in the two directions of the supply chain, and the constraints and behavioral logic faced by focal firms in the upstream and the downstream supply chain are different [8,9]. Does this lead to the asymmetry of the spillover effect of innovation in the upstream and the downstream supply chain? What is the mechanism? There is no existing literature to analyze and answer this. Answering these questions has significant implications for enriching the related literature on the spillover effect of innovation and focal firms, making it the core role of focal firms to promote supply chain innovation and lead the overall development of the supply chain.

In recent years, China’s economic and social development has faced a complex and severe situation with increased external pressure and more internal difficulties. The Chinese government has focused on achieving supply chain collaborative innovation by fostering focal firms. At the central government level, in 2020, the Ministry of Commerce and other departments issued the “Notice on Further Doing a Good Job in Supply Chain Innovation and Application Pilot Work”, focusing on cultivating focal firms in the supply chain and promoting supply chain collaborative innovation. In 2022, the Ministry of Industry and Information Technology released the first list of key industrial and supply chain focal firms, covering key fields such as integrated circuits and new energy vehicles. In 2024, the Ministry of Industry and Information Technology and other departments issued the “Guiding Opinions on Deepening the Development of Industrial Supply Chain Collaborative Cooperation”, proposing to establish a supply chain collaborative innovation mechanism led by focal firms and promoting the integrated development of large and small enterprises. At the local government level, each region has successively released the list of its key industrial and supply chain and focal firms. The measures and data of the Chinese government in the supply chain and focal firms have provided research opportunities for examining the bidirectional innovation spillovers from focal firms to upstream and downstream supply-chain firms.

Based on this, this paper uses the supply chain portfolio screened from the key industrial chains and focal firms announced by Chinese governments from 2009 to 2024 as the research sample. Firstly, it examines the bidirectional innovation spillovers from focal firms to upstream and downstream supply-chain firms. The research results show that there are the bidirectional innovation spillovers from focal firms to supply-chain firms, and the innovation spillover effect is asymmetric; that is, compared with the upstream supply chain, the innovation spillover from focal firms to supply-chain firms in the downstream supply chain is stronger. Secondly, it explores the mechanism of the innovation spillover effect of focal firms on supply-chain firms in the upstream supply chain and the downstream supply chain respectively. The research results show that focal firms can enhance the innovation level of upstream firms through the cooperative innovation mechanism, and can enhance the innovation level of downstream firms through the knowledge spillover mechanism.

The possible marginal contributions of this paper mainly include the following aspects. Firstly, it enriches the related research on innovation in supply-chain spillover. The previous literature has confirmed that the supply chain is an important channel for innovation spillover [5], and has discovered that there is innovation spillover among the nodes of the supply chain [10], but little literature has examined the bidirectional innovation spillovers from focal firms to supply-chain firms, especially the directional difference in the innovation spillover effect; that is, the spillover effect of focal firms on innovation in supply-chain firms in the upstream supply chain, while the downstream supply chain has different intensities. This paper further supplements the related research on innovation in supply-chain spillover. Secondly, it provides empirical evidence of the leading role of focal firms in the supply chain. The previous literature believes that focal firms can drive the overall development of the supply chain by leveraging their advantages such as technology, funds, and markets [6,7]. This paper uses the research opportunities provided by the key industrial chains and focal firms announced by Chinese governments from an innovation spillover perspective to provide empirical evidence of the leading role of focal firms and further explores its mechanism.

2. Literature Review and Theoretical Hypotheses

2.1. Literature Review

2.1.1. Spillover Effects of Innovation

First, regarding the entities involved in innovation spillover, existing studies include macro, meso and micro entities. From the macro perspective, innovation spillover mainly involves countries, regions and cities. For instance, developed countries can achieve innovation spillover to developing countries through means such as technical assistance, trade rules, and industrial transfer [11]. From the meso perspective, innovation spillover mainly involves industries and supply chains. For example, due to the existence of inter-industry correlation and leading demonstration effects, specific industries will actively or passively draw innovation resources from related industries [12]. Supply chains are an important channel for knowledge flow, and they can achieve innovation spillover internally through market demand and technology traction [13,14]. From the micro perspective, innovation spillover mainly involves enterprises, governments and non-governmental organizations. For instance, enterprises achieve a cross-enterprise flow of knowledge and technology through competition and cooperative relationships [2]. Under the concept of integrated research and production, enterprises and universities have extensive interactions in terms of technology transfer and diffusion [3].

Second, regarding the driving factors of innovation spillover, existing studies include exogenous and endogenous driving factors, and their combined effect influences the innovation spillover of entities. From the exogenous factors, changes in the constraints or support of elements will affect innovation spillover. For example, geographical proximity can accelerate the scale and efficiency of element flow and promote innovation spillover among entities [15]. A high correlation between industries and products can reduce the cost and loss of element flow and enhance the level of innovation spillover among entities [12]. Government industrial policies, special subsidies and targeted purchases will stimulate the innovation enthusiasm of entities, reduce the risk of innovation activities, and promote the innovation spillover of related entities [4]. From the endogenous factors, subjective motivations such as consolidating position, enhancing influence, and improving efficiency will affect innovation spillover [16]. For example, optimizing the corporate governance structure can improve the organization’s ability to recognize, absorb, transform and apply knowledge, and enhance the innovation spillover level of entities [17]. The differences and complementarities in institutional mechanisms and organizational structures between state-owned enterprises and private enterprises will affect the scale and efficiency of innovation spillover [18]. The effectiveness of innovation investment decisions not only affects their own innovation output level but also the innovation spillover effect [19].

2.1.2. Leading Role of Focal Firms in the Supply Chain

The supply chain is a chain-like division and cooperation relationship formed by numerous nodes of enterprises, based on certain technical and economic logic and spatial layout relationships [20]. The focal firms in the supply chain, with its advantages in technology, funds and markets, can integrate internal and external resources, connect upstream and downstream firms, and exert strong control and influence, leading the overall development of the supply chain [6]. The current literature holds that the leading role of focal firms in the supply chain mainly includes three aspects: technology, resources and contracts. From the technical perspective, the focal firms can not only guide the direction of technological progress and resource investment, but also lead the formulation and maintenance of industry, product rules and standards [21]. Further, the focal firms can output technological paradigms through interaction with supply-chain firms, thereby achieving innovation spillover and multiplication [22]. From the resource perspective, the focal firms can not only optimize supply-chain resource allocation through the platform to enhance the market response ability of supply-chain firms [23], but also improve the financial situation of supply-chain firms through supply-chain finance and trade collaboration, and strengthen the overall information transparency of the supply chain [24]. From the contractual perspective, the focal firms can significantly enhance the market competitiveness and growth potential of supply-chain firms through aggregation [25], and coordinate the procurement, research and development, production and sales of all links of the supply chain through supplier and customer management to improve supply-chain efficiency and resilience [26], and further enhance the consistency of the supply chain through merger and acquisition behavior [27].

2.1.3. Comment

Although the spillover effects of innovation and the leading role of focal firms in the supply chain have received extensive attention from the academic community, there are still some research gaps. First, there is a lack of evidence on the bidirectional innovation spillovers from focal firms to supply-chain firms. The previous literature has confirmed that the supply chain is an important channel for innovation spillover. However, few studies have examined the bidirectional innovation spillovers from focal firms to supply-chain firms, especially the directional difference in the innovation spillover effect. Second, the mechanism of the bidirectional innovation spillovers from focal firms to supply-chain firms is not well revealed. The previous literature has explored the driving factors of innovation spillover from aspects such as endogeneity and exogeneity. However, there is a lack of more-in-depth analysis and verification of the mechanism of the bidirectional innovation spillovers from focal firms to supply-chain firms. Therefore, this paper examines the bidirectional innovation spillover from focal firms to supply-chain firms and its mechanism, in order to enrich the relevant literature and provide empirical evidence and theoretical support for focal firms to play a core role, promote supply chain innovation, and lead the overall development of the supply chain.

2.2. Theoretical Hypotheses

The focal firm in the supply chain, as the core node of the supply chain, can generate significant spillover effects on innovation in supply-chain firms through technology leadership, resource integration, and contract coordination. Firstly, from the perspective of technology leadership, the focal firm dominates the output of technical standards and paradigms, and will require upstream and downstream supply-chain firms to adapt to its technical framework. According to the Global Value Chain (GVC) theory [28], the focal firm plays the role of a “technical architect” in the supply chain, and its technology selection directly defines the technical trajectory of the supply chain, thereby promoting technology diffusion and forcing upstream and downstream supply-chain firms to carry out adaptive innovation. At the same time, the technology leadership of the focal firm also lowers the innovation threshold of the supply-chain firms [29], further amplifying the spillover effect of the focal firm on innovation in the supply-chain firms. Secondly, from the perspective of resource integration, the focal firm’s resource advantages enable it to configure innovative resources through the supply chain to promote knowledge diffusion. The focal firm can not only alleviate the financial constraints of the supply-chain firms through supply-chain finance, R&D subsidies, etc. [30], but can also reduce the innovation risks of the supply-chain firms through information sharing such as demand forecasting and market trend analysis, and can also integrate supply-chain resources through platform tools [13], improving the innovation efficiency of the supply-chain firms and ultimately reducing innovation costs and enhancing their innovation willingness and ability, promoting the spillover effect of the focal firm on innovation in the supply-chain firms. Thirdly, from the perspective of contract coordination, the focal firm can, through contract design and institutional arrangements, force or guide the supply-chain firms to participate in innovation activities, reduce opportunistic behavior in the supply-chain firms in the innovation process [31], and promote knowledge spillover. For example, the focal firm can share innovation risks and benefits with the supply-chain firms through joint R&D methods such as technology alliances and co-building laboratories, achieving collaborative innovation [32]. The focal firm can also improve the stability of innovation activities by signing supplier or customer management agreements, including quality certification, environmental standards, long-term order commitments, etc. [10]. Therefore, the spillover effect of innovation exists widely in the relationship between the focal firm and the supply-chain firms. The following will further elaborate on the bidirectional innovation spillovers from focal firms to supply-chain firms, respectively, for the upstream supply chain and the downstream supply chain.

First, the innovation spillover from focal firms to upstream firms: The focal firm in the upstream supply chain is closer to the primary products and the technical foundation. Therefore, the focal firm at the demand side faces higher levels of innovation risks and resource dependence. They must promote knowledge progress and technological iteration in a reliable and stable manner. At this time, collaborative innovation is an important mechanism for driving innovation spillover [32]. According to the social network theory, each supply-chain node enterprise’s innovation resources and innovation decisions will have a significant impact on the entire supply chain, and the close cooperation among supply-chain node enterprises is a strategic way to obtain innovation resources and make collaborative innovation decisions [5]. On the one hand, the cooperative innovation between the focal firm and the upstream firms can share innovation risks and change the risk distribution of the supply chain, thereby reducing the uncertainty of innovation activities and promoting the spillover effect [30]. On the other hand, the cooperative innovation between the focal firm and the upstream firms can optimize supply-chain resource allocation, alleviate the resource dependence of a single node on innovation activities, and maximize the utilization of supply-chain technological accumulation, motivating the efforts of supply-chain node enterprises, and promoting the innovation spillover from focal firms to upstream firms.

Second, the innovation spillover from focal firms to downstream firms: The focal firm in the downstream supply chain is closer to the final products and the end markets. Therefore, the focal firm at the supply side must promptly respond to product demands and market feedback, accurately grasping the direction of knowledge advancement and technological iteration. At this time, knowledge spillover is an important mechanism driving innovation spillover [33]. Audretsch and Feldman (1996) proposed the theory of knowledge spillover and the distribution of innovation space, which holds that geographical distance is a key obstacle to knowledge and technology spillover [34]. That is, the spillover effect of knowledge and technology will weaken as the geographical distance increases. This is because knowledge and technology spread in space will experience lag, attenuation, and distortion, especially for tacit knowledge and technology that are difficult to textualize, which are crucial for innovation activities [35]. Therefore, by maintaining a certain geographical distance from downstream firms, focal firms can achieve a high-quality flow of knowledge and technology, which is the key to innovation spillover. On the contrary, a large geographical distance between focal firms and downstream firms will lead to a decrease in the frequency of interaction between the two parties, thereby limiting knowledge spillover and ultimately weakening the innovation spillover from focal firms to upstream firms.

However, it should be noted that focal firms occupy different ecological positions in the two directions of the supply chain, and the constraints and behavioral logic faced by focal firms in the upstream and the downstream supply chain are different [8,9]. This leads to a directional difference in the bidirectional innovation spillover from focal firms to supply-chain firms; that is, the spillover effect of innovation from focal firms to upstream and downstream firms has different intensities. Further, the intensity of the innovation spillover effect from focal firms to supply-chain firms in different directions mainly depends on the control power of focal firms, the motivation for resource investment, and the transmission of knowledge and information. In these three aspects, the focal firms can more effectively drive the innovation activities of downstream firms; that is, compared with the upstream supply chain, the innovation spillover from focal firms to supply-chain firms in the downstream supply chain is stronger.

Specifically, first, from the perspective of control power, focal firms usually have core product technologies and can directly define product specifications, service standards, marketing strategies, and even business models, having significant bargaining power and market dominance over downstream firms, thus forcing or guiding downstream firms to carry out corresponding innovations to meet their requirements. In the upstream supply chain, focal firms are at the demand side, and the innovation requirements of focal firms for upstream firms are mainly technical compatibility [31], which limits the intensity of innovation spillover. Second, from the perspective of the motivation for resource investment, focal firms have a strong motivation to improve the sales efficiency and service quality of downstream firms, because this directly affects their own products’ market performance and brand reputation [13]. Therefore, focal firms are more willing to actively invest resources in downstream firms to promote innovation [30]. This direct and proactive resource investment is an important driving force for innovation spillover to focal firms in the downstream supply chain, and its intensity and scope are usually much greater than similar investments by focal firms in upstream firms. Third, from the perspective of the transmission of knowledge and information, the information flow in the downstream supply chain is mainly downward information such as market demand and consumer preferences, and focal firms have the motivation and ability to ensure that downstream firms can understand, absorb, and apply this knowledge information [2]. With the high scale and efficiency of knowledge information transmission, the innovation spillover effect is also stronger. In contrast, the information flow in the upstream supply chain is mainly upward information such as raw materials and basic technologies, and the influence of focal firms is more about assessment and selection, and the intensity of innovation spillover is accordingly reduced.

Based on the above analysis, this paper proposes the following research hypotheses:

Hypothesis 1.

There are bidirectional innovation spillovers from focal firms to supply-chain firms.

Hypothesis 2.

Compared with the upstream supply chain, the innovation spillover from focal firms to supply-chain firms in the downstream supply chain is stronger.

3. Data and Methodology

3.1. Data

This paper takes the supply-chain portfolio of upstream firm–focal firm–downstream firm as the research sample. As shown in Figure 1, the sample screening process is as follows: (1) The initial sample consists of Chinese A-share listed companies on the Shanghai and Shenzhen stock exchanges from 2009 to 2024. (2) Financial, ST, and data-missing or abnormal samples are excluded. (3) Focal firms and supply-chain firms are selected based on the list of key industrial chains and focal firms announced by Chinese governments at all levels. For example, according to the “Provincial Key Industrial Chains” published by the Government of Shanxi Province and “List of Key Focal Firms in Key Industrial Chains of Shanxi Province” published by the Shanxi Provincial Department of Industry and Information Technology, Shanxi Lu’an Solar Technology Co., Ltd. is the focal firm of the photovoltaic industrial chain in Shanxi Province. Further, through the control relationship disclosed on the company’s official website and annual reports, it is confirmed that this enterprise is a subsidiary of the listed company Lu’an Energy (stock code: 601699). Finally, based on the above documents, the focal firm and other supply-chain firms of the photovoltaic industrial chain in Shanxi Province are screened out. (4) Considering that the focal firm and supply-chain firms may have multiple corresponding relationships, we have constructed a supply-chain portfolio of upstream firm–focal firm–downstream firm. For example, if focal firm L has three upstream firms (F1 to F3) and one downstream firm (B), then three supply-chain portfolios can be formed: F1—L—B, F2—L—B, and F3—L—B. (5) To facilitate data acquisition, only supply-chain portfolios where both the focal firm and supply-chain firms are listed companies are retained. Furthermore, regarding the selection criteria, the local areas are based on the standards set by the central government. As the release of the relevant list is within the past 5 years, and the policies and documents of the central government in China have not undergone any substantive changes or updates, this does not affect the sample selection process of this paper. Therefore, criteria are consistent across provinces and years. Ultimately, this paper obtains 7136 supply-chain portfolios, as detailed in

Table 1. Sample statistics.

Sample	Upstream Sample	Downstream Sample	Entire Sample
Definition	Upstream Firm– Focal Firm	Focal Firm– Downstream Firm	Upstream Firm–Focal Firm–Downstream Firm
N	2249	4887	7136
%	31.52	68.48	100

. This paper also applies a 1% Winsorization to continuous indicators to control the influence of extreme values. The list of focal firms is manually compiled from the official websites of Chinese governments at all levels and listed companies, while the remaining data are sourced from the CSMAR and WIND databases. Stata 18.5 is used as the statistical software in this paper.

3.2. Model Specification and Variables

Drawing on the existing literature [36], this paper constructs the following model to examine the bidirectional innovation spillovers from focal firms to upstream and downstream supply-chain firms:

$$\begin{array}{c}\mathrm{I}\mathrm{C}\_\mathrm{C}\mathrm{H}\mathrm{A}\mathrm{I}{\mathrm{N}}_{\mathrm{i},\mathrm{t}}={\mathrm{\beta }}_0+{\mathrm{\beta }}_1\mathrm{I}\mathrm{C}\_\mathrm{F}\mathrm{O}\mathrm{C}\mathrm{A}{\mathrm{L}}_{\mathrm{i},\mathrm{t}}+\mathrm{\beta }\mathrm{C}\mathrm{O}\mathrm{N}\mathrm{T}\mathrm{R}\mathrm{O}{\mathrm{L}}_{\mathrm{i},\mathrm{t}}\\ +\mathrm{Y}\mathrm{E}\mathrm{A}{\mathrm{R}}_{\mathrm{i},\mathrm{t}}+\mathrm{I}\mathrm{N}{\mathrm{D}}_{\mathrm{i},\mathrm{t}}+\mathrm{F}\mathrm{I}\mathrm{R}{\mathrm{M}}_{\mathrm{i},\mathrm{t}}+{\mathrm{\epsilon }}_{\mathrm{i},\mathrm{t}}\end{array}$$

(1)

Among them, the dependent variable IC_CHAIN represents innovation in supply-chain firms, while the independent variable IC_FOCAL represents the innovation of focal firms. Drawing on the existing literature [37], this paper measures the level of innovation in supply-chain firms and focal firms by using the logarithmized number of patent applications [4]. The larger the value of IC_CHAIN and IC_FOCAL, the higher the innovation level. CONTROL represents the set of control variables. Additionally, the model also controls for the fixed effects of the year (YEAR), industry (IND), and firm-specific (FIRM) fixed effects, and the specific definitions are provided in

Table 2. Variable definitions.

Variable	Definition
IC_CHAIN	The natural logarithm of one plus the number of patent applications from supply-chain firms.
IC_FOCAL	The natural logarithm of one plus the number of patent applications from focal firms.
SIZE	The natural logarithm of total assets.
LEV	The ratio of total liabilities to total assets.
ROE	Return on equity defined as net profit divided by shareholders’ equity.
R&D	The natural logarithm of one plus the amount of research and development expenditure.
FATA	The ratio of fixed assets to total assets.
GS	The natural logarithm of one plus the number of government subsidies.
HHI	The sales-based Herfindahl–Hirschman index of a firm.
AGE	The natural logarithm of a firm’s age since the year it was established.
YEAR	Fixed effects of the year.
IND	Fixed effects of the industry.
FIRM	Firm-specific fixed effects.

.

3.3. Descriptive Statistics

Table 3. Descriptive statistics results.

Variable	N	Mean	Std	Min	Median	Max
IC_CHAIN	7136	3.797	2.376	0	3.714	9.610
IC_FOCAL	7136	2.579	1.803	0	2.708	8.994
SIZE	7136	23.500	1.961	20.230	23.320	30.250
LEV	7136	0.455	0.219	0.009	0.466	0.937
ROE	7136	0.071	0.097	−0.355	0.066	0.343
R&D	7136	18.920	2.064	12.360	18.360	23.650
FATA	7136	0.131	0.157	0	0.069	0.638
GS	7136	17.330	2.809	0	17.510	22.240
HHI	7136	0.154	0.161	0.014	0.096	1
AGE	7136	2.779	0.424	0	2.833	4.043

presents the descriptive statistical results of the main variables in this study. As shown in Table 3, the mean, median, and standard deviation of IC_CHAIN are 3.797, 3.714, and 2.376 respectively, which indicate that there are significant differences in innovation performance across supply-chain firms; the mean, median, and standard deviation of IC_FOCAL are 2.579, 2.708, and 1.803 respectively, also reflecting the differences in innovation performance in different focal firms. By further comparing the descriptive statistical results of IC_CHAIN and IC_FOCAL, it can be seen that the standard deviation of IC_CHAIN is greater than that of IC_FOCAL, which indicates that the innovation spillover effect of focal firms may have different intensities for different supply-chain firms, thereby causing the standard deviation of innovation from supply-chain firms to increase. This provides a research space for this study to further examine the differences in the innovation spillover effects between the upstream sample and the downstream sample. The statistical results of other variables are basically consistent with the existing literature, reflecting the reliability of data collection and organization in this study.

4. Empirical Results

4.1. Baseline Regression

Table 4. Baseline regression results.

Variable	(1)	(2)	(3)	(4)
	Entire Sample	Entire Sample	Upstream Sample	Downstream Sample
IC_FOCAL	0.161 ***	0.107 ***	0.038 ***	0.148 ***
	(13.954)	(12.249)	(2.871)	(13.046)
SIZE		0.770 ***	0.638 ***	0.801 ***
		(47.438)	(21.948)	(40.204)
LEV		−0.084	−0.018	−0.095
		(−1.019)	(−0.134)	(−0.912)
ROE		0.230	−0.001	0.388 *
		(1.449)	(−0.003)	(1.927)
R&D		0.105 ***	0.195 ***	0.075 ***
		(10.264)	(10.472)	(6.070)
FATA		−0.642 ***	−0.746 ***	−0.711 ***
		(−4.845)	(−3.249)	(−4.437)
GS		0.036 ***	0.037 ***	0.038 ***
		(5.535)	(3.546)	(4.627)
HHI		0.649 ***	0.386	0.801 ***
		(3.666)	(0.949)	(3.969)
AGE		0.028	−0.060	0.025
		(0.633)	(−0.755)	(0.475)
Constant	0.143	−18.980 ***	−16.690 ***	−19.380 ***
	(0.307)	(−40.043)	(−24.760)	(−26.279)
Year/Ind/Firm FE	Yes	Yes	Yes	Yes
N	7136	7136	2249	4887
R2	0.5555	0.7458	0.7804	0.7461
Diff			0.000 ***
Beta	12.18%	8.13%	3.27%	10.69%

Note: t-statistics are in parentheses; * p < 0.1, *** p < 0.01; Diff represents the p-value for the difference in inter-group coefficients, which is calculated based on the estimation results of the non-correlation test.

presents the regression results of the bidirectional innovation spillovers from focal firms to upstream and downstream supply-chain firms. Columns (1) and (2) represent the regression results for the entire sample. In column (1), only the year, industry, and firm-specific fixed effects were controlled for. In column (2), additional control variables were included. From the regression results of columns (1) and (2), it can be seen that the regression coefficients of IC_FOCAL are all significantly positive, verifying Hypothesis H1, which indicates that there are bidirectional innovation spillovers from focal firms to supply-chain firms. At the same time, from the perspective of economic significance, from the Beta of IC_FOCAL in columns (1) and (2), when only the year, industry, and firm-specific fixed effects were controlled for, focal firms could enhance innovation in supply-chain firms by approximately 12.18% on average. At this point, for every one standard deviation increase in IC_FOCAL, the logarithmized innovation of the supply-chain firms increases by approximately 0.290. However, after controlling for other influencing factors and the effects of the year, industry, and firm-specific fixed effects, focal firms could enhance innovation in supply-chain firms by approximately 8.13% on average. At this point, for every one standard deviation increase in IC_FOCAL, the logarithmized innovation of the supply-chain firms increases by approximately 0.193. Additionally, the regression results of the control variables are largely consistent with the previous literature.

Furthermore, this paper divides the research samples into “upstream firm–focal firm” for upstream samples and “focal firm–downstream firm” for downstream samples. Using model (1), it examines whether the innovation spillover effect is asymmetric between upstream and downstream samples. Column (3) is the upstream sample, and the regression coefficient of IC_FOCAL is significantly positive at the 1% level. Column (4) is the downstream sample, and the regression coefficient of IC_FOCAL is also significantly positive at the 1% level. However, the inter-group differences in columns (3) and (4) are significant at the 1% level, indicating that the innovation spillover effect of focal firms to supply-chain firms is significantly different in the upstream and downstream samples. At the same time, from an economic perspective, as can be seen from the Beta of IC_FOCAL in columns (3) and (4), after controlling for the influence of other factors and the year, industry, and firm-specific fixed effects, in the upstream sample, focal firms can enhance innovation in supply-chain firms by approximately 3.27% on average, while in the downstream sample, focal firms can enhance innovation in supply-chain firms by approximately 10.69% on average. This indicates that, compared with the upstream supply chain, the innovation spillover from focal firms to supply-chain firms in the downstream supply chain is stronger, verifying Hypothesis H2.

4.2. Robustness Tests

4.2.1. Endogeneity Test

From a theoretical perspective, there may be endogeneity issues in the bidirectional innovation spillovers from focal firms to upstream and downstream supply-chain firms. This endogeneity mainly stems from reverse causality. For example, supply-chain firms can request focal firms to produce the intermediate products they need according to their own requirements and standards. This process may involve technological spillovers from supply-chain firms to focal firms [10]. Based on this, this paper sets the following two instrumental variables for robustness testing:

First, this paper uses the pilot on innovative cities as the policy shock to identify the causal effect of the focal firms to supply-chain firms on innovation. The National Development and Reform Commission and the Ministry of Science and Technology of China successively approved 78 national pilot innovative cities in 2008 and thereafter. Theoretically, the participation of focal firms in the pilot on innovative cities is conducive to improving their innovation level, but it will not directly affect supply-chain firms, thus meeting the requirements for the relevance and exogeneity of the instrumental variable. Therefore, this paper uses whether focal firms are based in a pilot city as the instrumental variable and re-estimates using model (1). Moreover, it should be pointed out that this paper excludes the samples where the focal firms and supply-chain firms belong to the same city. As shown in columns (1) to (3) of

Table 5. IV test results.

Variable	(1)	(2)	(3)	(4)	(5)	(6)
	IV1			IV2
	Entire Sample	Upstream Sample	Downstream Sample	Entire Sample	Upstream Sample	Downstream Sample
IC_IV	0.175 ***	0.091 *	0.150 ***	0.117 ***	0.016	0.126 ***
	(4.309)	(2.778)	(2.879)	(4.845)	(0.424)	(4.064)
SIZE	0.776 ***	0.645 ***	0.812 ***	0.779 ***	0.640 ***	0.815 ***
	(47.330)	(22.186)	(39.998)	(47.049)	(21.401)	(39.983)
LEV	−0.092	−0.021	−0.109	−0.110	−0.030	−0.135
	(−1.097)	(−0.161)	(−1.025)	(−1.308)	(−0.222)	(−1.260)
ROE	0.264 *	0.071	0.299	0.237	0.008	0.318
	(1.650)	(0.293)	(1.464)	(1.471)	(0.031)	(1.539)
R&D	0.106 ***	0.194 ***	0.073 ***	0.109 ***	0.204 ***	0.080 ***
	(10.186)	(10.419)	(5.809)	(10.405)	(10.509)	(6.266)
FATA	−0.662 ***	−0.695 ***	−0.806 ***	−0.695 ***	−0.762 ***	−0.804 ***
	(−4.934)	(−3.017)	(−4.941)	(−5.149)	(−3.276)	(−4.881)
GS	0.037 ***	0.036 ***	0.041 ***	0.037 ***	0.036 ***	0.040 ***
	(5.696)	(3.521)	(4.977)	(5.537)	(3.438)	(4.774)
HHI	0.656 ***	0.363	0.813 ***	0.591 ***	0.447	0.736 ***
	(3.670)	(0.893)	(3.959)	(3.180)	(0.995)	(3.485)
AGE	0.034	−0.067	0.046	0.044	−0.083	0.053
	(0.766)	(−0.842)	(0.863)	(1.001)	(−1.029)	(0.987)
Constant	−19.100 ***	−16.890 ***	−19.680 ***	−19.410 ***	−17.000 ***	−19.650 ***
	(−39.881)	(−24.887)	(−26.254)	(−35.803)	(−23.539)	(−20.076)
Year/Ind /Firm FE	Yes	Yes	Yes	Yes	Yes	Yes
N	5708	1799	3909	6953	2153	4800
R2	0.7411	0.7804	0.7376	0.7415	0.7744	0.7378
Diff		0.008 ***			0.044 **
Beta	10.18%	2.04%	9.16%	4.80%	0.71%	5.03%

Note: t-statistics are in parentheses; * p < 0.1, ** p < 0.05, *** p < 0.01.

, the regression coefficients of IC_IV are all significantly positive, and the differences between groups in columns (2) and (3) are significant at the 1% level. Additionally, the Beta results are also consistent with the baseline regression, still supporting the hypotheses of this paper.

Secondly, this paper takes the average innovation level of other sample focal firms in the same industry and the same year, excluding the sample focal firms themselves, as the instrumental variable, and model (1) is re-estimated. As shown in columns (4) to (6) of Table 5, the regression coefficient of IC_IV is significantly positive in the entire sample and the downstream sample, positive but not significant in the upstream sample, and the differences between groups in columns (5) and (6) are significant at the 5% level. At the same time, the Beta results are consistent with the baseline regression and still support the hypotheses of this paper.

4.2.2. The Lagging Effect of Innovation Spillovers

Previous studies have confirmed that innovation spillovers have a lagging effect [5]. Therefore, in order to further enhance the reliability of the research conclusions, this paper sets the variable of the focal firm’s innovation to lag by one period and two periods, and uses model (1) to re-estimate. Columns (1) to (3) of

Table 6. Lagging effect test results.

Variable	(1)	(2)	(3)	(4)	(5)	(6)
	LAG1			LAG2
	Entire Sample	Upstream Sample	Downstream Sample	Entire Sample	Upstream Sample	Downstream Sample
IC_LAG	0.085 ***	0.041 **	0.107 ***	0.108 ***	0.010	0.140 ***
	(7.352)	(2.206)	(7.346)	(7.626)	(0.387)	(8.126)
SIZE	0.815 ***	0.702 ***	0.852 ***	0.826 ***	0.758 ***	0.839 ***
	(41.004)	(20.389)	(34.134)	(34.139)	(16.931)	(28.109)
LEV	0.015	0.068	−0.081	0.185	−0.312	0.200
	(0.141)	(0.432)	(−0.602)	(1.465)	(−1.503)	(1.249)
ROE	0.014	−0.480 *	0.250	−0.207	−0.700 **	−0.085
	(0.075)	(−1.714)	(1.022)	(−0.933)	(−1.975)	(−0.298)
R&D	0.091 ***	0.187 ***	0.055 ***	0.104 ***	0.211 ***	0.072 ***
	(7.685)	(8.653)	(3.818)	(7.336)	(7.840)	(4.183)
FATA	−0.276 *	−0.693 **	−0.240	−0.192	−0.197	−0.127
	(−1.701)	(−2.510)	(−1.190)	(−0.998)	(−0.584)	(−0.535)
GS	0.055 ***	0.050 ***	0.062 ***	0.036 ***	0.041 ***	0.042 ***
	(6.671)	(4.135)	(5.725)	(3.751)	(2.807)	(3.384)
HHI	−0.304	0.450	−0.272	1.344 ***	0.795	1.522 ***
	(−1.321)	(0.974)	(−0.992)	(5.186)	(1.464)	(5.054)
AGE	0.065	0.035	0.075	0.203 ***	0.057	0.227 **
	(1.141)	(0.336)	(1.046)	(2.783)	(0.410)	(2.560)
Constant	−19.910 ***	−18.590 ***	−20.830 ***	−20.220 ***	−19.160 ***	−20.850 ***
	(−32.533)	(−22.753)	(−16.381)	(−28.015)	(−18.512)	(−15.536)
Year/Ind /Firm FE	Yes	Yes	Yes	Yes	Yes	Yes
N	4498	1424	3074	3442	1060	2382
R2	0.7951	0.8470	0.7824	0.7887	0.8288	0.7820
Diff		0.006 ***			0.000 ***
Beta	5.60%	2.73%	7.09%	6.80%	0.60%	8.88%

Note: t-statistics are in parentheses; * p < 0.1, ** p < 0.05, *** p < 0.01.

show the regression results of the lagged one-period variable of the focal firm’s innovation. The regression coefficients of IC_LAG are all significantly positive, and the differences between columns (2) and (3) are significant at the 1% level. At the same time, the Beta results are also consistent with the baseline regression and still support hypotheses H1 and H2. Columns (4) to (6) of Table 6 show the regression results of the lagged two-period variable of the focal firm’s innovation. The regression coefficients of IC_LAG are significantly positive in the entire sample and the downstream sample, and positive but insignificant in the upstream sample. The differences between columns (5) and (6) are also significant at the 1% level, and the Beta results are also consistent with the baseline regression and still support the hypotheses of this paper.

4.2.3. Alternative Indicator Test

This paper re-measures the level of innovation in supply-chain firms and focal firms by using the natural logarithm of one plus the number of invention patent applications, and then regresses using model (1). As shown in

Table 7. Alternative indicator test results.

Variable	(1)	(2)	(3)	(4)
	Entire Sample	Entire Sample	Upstream Sample	Downstream Sample
IC_FOCAL	0.186 ***	0.121 ***	0.074 ***	0.159 ***
	(14.804)	(12.798)	(5.131)	(12.990)
SIZE		0.742 ***	0.659 ***	0.750 ***
		(47.021)	(22.429)	(39.226)
LEV		−0.251 ***	−0.371 ***	−0.175 *
		(−3.114)	(−2.803)	(−1.739)
ROE		0.072	−0.200	0.279
		(0.464)	(−0.818)	(1.444)
R&D		0.132 ***	0.209 ***	0.107 ***
		(13.202)	(11.107)	(8.937)
FATA		−0.422 ***	−0.209	−0.564 ***
		(−3.273)	(−0.900)	(−3.664)
GS		0.046 ***	0.052 ***	0.046 ***
		(7.226)	(5.020)	(5.860)
HHI		0.620 ***	−0.447	0.860 ***
		(3.603)	(−1.088)	(4.431)
AGE		0.063	0.091	0.023
		(1.498)	(1.144)	(0.457)
Constant	−0.126	−19.280 ***	−17.990 ***	−18.970 ***
	(−0.277)	(−41.819)	(−26.425)	(−26.789)
Year/Ind/Firm FE	Yes	Yes	Yes	Yes
N	7136	7136	2249	4887
R2	0.5267	0.7340	0.7697	0.7353
Diff			0.000 ***
Beta	13.18%	8.59%	5.95%	10.67%

Note: t-statistics are in parentheses; * p < 0.1, *** p < 0.01.

, the regression coefficients of IC_FOCAL are all significantly positive, and the differences between the groups in column (3) and column (4) are significant at the 1% level. At the same time, the Beta results are also consistent with the baseline regression and still support the hypotheses of this paper.

4.2.4. Poisson Regression

This paper takes into account the characteristics of innovative data, which have a high frequency of zero values and a discrete distribution. Therefore, it re-applies Poisson regression. As shown in

Table 8. Poisson regression results.

Variable	(1)	(2)	(3)	(4)
	Entire Sample	Entire Sample	Upstream Sample	Downstream Sample
IC_FOCAL	0.041 ***	0.029 ***	0.009	0.044 ***
	(11.292)	(7.947)	(1.595)	(9.131)
SIZE		0.231 ***	0.153 ***	0.259 ***
		(30.977)	(10.165)	(28.082)
LEV		−0.017	0.023	−0.029
		(−0.448)	(0.354)	(−0.617)
ROE		0.046	0.029	0.027
		(0.653)	(0.238)	(0.298)
R&D		−0.005	0.036 ***	−0.016 ***
		(−1.034)	(3.532)	(−3.001)
FATA		−0.235 ***	−0.276 **	−0.252 ***
		(−3.791)	(−2.266)	(−3.389)
GS		0.009 ***	0.008	0.009 **
		(3.025)	(1.526)	(2.224)
HHI		0.244 ***	0.289	0.267 ***
		(3.326)	(1.547)	(3.246)
AGE		0.015	0.002	0.019
		(0.772)	(0.056)	(0.813)
Constant	0.002	−5.174 ***	−3.873 ***	−5.689 ***
	(0.009)	(−19.009)	(−9.709)	(−12.641)
Year/Ind/Firm FE	Yes	Yes	Yes	Yes
N	7136	7136	2249	4887
R2	0.1858	0.2458	0.2074	0.2699
Diff			0.000 ***

Note: Z-statistics are in parentheses; ** p < 0.05, *** p < 0.01.

, the regression coefficients of IC_FOCAL are significantly positive in the entire and downstream samples. Meanwhile, the regression coefficient of IC_FOCAL is positive but not significant in the upstream sample. And the differences between the groups in column (3) and column (4) are significant at the 1% level. The results still support the hypotheses of this paper.

5. Mechanism Analysis

The analysis results in the above text show that there are bidirectional innovation spillovers from focal firms to supply-chain firms. Moreover, compared with the upstream supply chain, the innovation spillover from focal firms to supply-chain firms in the downstream supply chain is stronger. In the theoretical analysis, this paper argues that the main mechanism by which focal firms generate innovation spillover effects in the upstream supply chain is collaborative innovation, and the main mechanism by which they generate innovation spillover effects in the downstream supply chain is knowledge spillover. Based on model (1), this paper constructs models (2) and (3) in addition to test the above mechanisms, and to provide evidential support for the hypotheses of this paper.

$$\begin{array}{c}\mathrm{M}\mathrm{e}{\mathrm{d}}_{\mathrm{i},\mathrm{t}}={\mathrm{\beta }}_0+{\mathrm{\beta }}_1\mathrm{I}\mathrm{C}\_\mathrm{F}\mathrm{O}\mathrm{C}\mathrm{A}{\mathrm{L}}_{\mathrm{i},\mathrm{t}}+\mathrm{\beta }\mathrm{C}\mathrm{O}\mathrm{N}\mathrm{T}\mathrm{R}\mathrm{O}{\mathrm{L}}_{\mathrm{i},\mathrm{t}}\\ +\mathrm{Y}\mathrm{E}\mathrm{A}{\mathrm{R}}_{\mathrm{i},\mathrm{t}}+\mathrm{I}\mathrm{N}{\mathrm{D}}_{\mathrm{i},\mathrm{t}}+\mathrm{F}\mathrm{I}\mathrm{R}{\mathrm{M}}_{\mathrm{i},\mathrm{t}}+{\mathrm{\epsilon }}_{\mathrm{i},\mathrm{t}}\end{array}$$

(2)

$$\begin{array}{c}\mathrm{I}\mathrm{C}\_\mathrm{C}\mathrm{H}\mathrm{A}\mathrm{I}{\mathrm{N}}_{\mathrm{i},\mathrm{t}}={\mathrm{\beta }}_0+{\mathrm{\beta }}_1\mathrm{I}\mathrm{C}\_\mathrm{F}\mathrm{O}\mathrm{C}\mathrm{A}{\mathrm{L}}_{\mathrm{i},\mathrm{t}}+{\mathrm{\beta }}_2\mathrm{M}\mathrm{e}{\mathrm{d}}_{\mathrm{i},\mathrm{t}}+\mathrm{\beta }\mathrm{C}\mathrm{O}\mathrm{N}\mathrm{T}\mathrm{R}\mathrm{O}{\mathrm{L}}_{\mathrm{i},\mathrm{t}}\\ +\mathrm{Y}\mathrm{E}\mathrm{A}{\mathrm{R}}_{\mathrm{i},\mathrm{t}}+\mathrm{I}\mathrm{N}{\mathrm{D}}_{\mathrm{i},\mathrm{t}}+\mathrm{F}\mathrm{I}\mathrm{R}{\mathrm{M}}_{\mathrm{i},\mathrm{t}}+{\mathrm{\epsilon }}_{\mathrm{i},\mathrm{t}}\end{array}$$

(3)

In model (2) and model (3), Med represents the mediating variable, including cooperative innovation (IC_CO1 and IC_CO2) and knowledge spillover (IC_DIS1 and IC_DIS2). This paper measures the degree of cooperative innovation (IC_CO1) by the number of patents jointly applied for by focal firms and supply-chain firms. Meanwhile, this paper measures the degree of cooperative innovation (IC_CO2) by the natural logarithm of one plus the number of research and development expenditures for the cooperating institutions or cooperative projects, which are disclosed in their annual report or on the company’s official website. The larger the IC_CO1 and the IC_CO2, the higher the level of cooperative innovation between focal firms and supply-chain firms. This paper measures the degree of knowledge spillover (IC_DIS1) by the spatial distance between focal firms and supply-chain firms. The larger the IC_DIS1, the greater the geographical distance between focal firms and supply-chain firms, and the lower the level of knowledge spillover from focal firms to supply-chain firms. The geographical distance is measured by the straight-line distance between the cities where focal firms and supply-chain firms are located. Meanwhile, this paper measures the degree of knowledge spillover (IC_DIS2) by technological proximity. This paper follows the approach from Jaffe (1986) [38] to measure technological proximity. The larger the IC_DIS2, the greater the level of knowledge spillover from focal firms to supply-chain firms.

• Spillover effect mechanism of upstream supply chain: cooperative innovation. As shown in columns (1) and (3) of Table 9, the regression coefficients of IC_FOCAL are all significantly positive. Further, by placing IC_FOCAL, IC_CO (IC_CO1 and IC_CO2), and IC_CHAIN into model (3), as shown in columns (2) and (4) of Table 9, the regression coefficients of IC_CO are all significantly positive, and the regression coefficients of IC_FOCAL are all still significantly positive and pass the Sobel test. This indicates that the mediating effect of cooperative innovation is significant; that is, focal firms can enhance the innovation level of upstream firms through the cooperative innovation mechanism.
• Spillover effect mechanism of downstream supply chain: knowledge spillover. As shown in column (1) of Table 10, the regression coefficient of IC_FOCAL is significantly negative. Further, by placing IC_FOCAL, IC_DIS1, and IC_CHAIN into model (3), as shown in column (2) of Table 10, the regression coefficient of IC_DIS1 is significantly negative, and the regression coefficient of IC_FOCAL is still significantly positive and passes the Sobel test. Meanwhile, as shown in column (3) of Table 10, the regression coefficient of IC_FOCAL is significantly positive. Further, by placing IC_FOCAL, IC_DIS2, and IC_CHAIN into model (3), as shown in column (4) of Table 10, the regression coefficient of IC_DIS2 is significantly positive, and the regression coefficient of IC_FOCAL is still significantly positive and passes the Sobel test. This indicates that the mediating effect of knowledge spillover is significant; that is, focal firms can enhance the innovation level of downstream firms through the knowledge spillover mechanism.

Table 9. Mechanism test results of upstream supply chain.

Variable	(1)	(2)	(3)	(4)
	IC_CO1		IC_CO2
	Med	IC_CHAIN	Med	IC_CHAIN
IC_FOCAL	0.031 **	0.034 **	0.094 ***	0.134 ***
	(2.199)	(2.570)	(10.974)	(12.261)
Med	−	0.141 ***	−	0.240 ***
	−	(6.948)	−	(4.196)
SIZE	0.351 ***	0.589 ***	0.674 ***	0.700 ***
	(11.489)	(19.873)	(42.623)	(36.463)
LEV	−0.814 ***	0.097	−0.013	−0.080
	(−5.914)	(0.742)	(−0.161)	(−0.790)
ROE	−0.913 ***	0.128	0.528 ***	0.791 ***
	(−3.592)	(0.532)	(3.420)	(4.082)
R&D	0.047 **	0.188 ***	0.107 ***	0.077 ***
	(2.400)	(10.213)	(10.681)	(6.470)
FATA	−0.551 **	−0.669 ***	−0.460 ***	−0.405 ***
	(−2.282)	(−2.940)	(−3.560)	(−2.624)
GS	0.044 ***	0.030 ***	0.031 ***	0.034 ***
	(4.066)	(2.966)	(4.860)	(4.388)
HHI	−0.991 **	0.525	0.703 ***	0.779 ***
	(−2.320)	(1.305)	(4.072)	(4.005)
AGE	0.240 ***	−0.094	0.002	0.004
	(2.884)	(−1.192)	(0.046)	(0.071)
Constant	−9.467 ***	−15.360 ***	−17.500 ***	−18.020 ***
	(−13.370)	(−22.135)	(−37.880)	(−25.376)
Year/Ind/Firm FE	Yes	Yes	Yes	Yes
N	2249	2249	2249	2249
R2	0.4053	0.7851	0.4382	0.7402
Sobel Test		0.000 ***		0.004 **

Note: t-statistics are in parentheses; ** p < 0.05, *** p < 0.01.

Table 9. Mechanism test results of upstream supply chain.

Variable	(1)	(2)	(3)	(4)
	IC_CO1		IC_CO2
	Med	IC_CHAIN	Med	IC_CHAIN
IC_FOCAL	0.031 **	0.034 **	0.094 ***	0.134 ***
	(2.199)	(2.570)	(10.974)	(12.261)
Med	−	0.141 ***	−	0.240 ***
	−	(6.948)	−	(4.196)
SIZE	0.351 ***	0.589 ***	0.674 ***	0.700 ***
	(11.489)	(19.873)	(42.623)	(36.463)
LEV	−0.814 ***	0.097	−0.013	−0.080
	(−5.914)	(0.742)	(−0.161)	(−0.790)
ROE	−0.913 ***	0.128	0.528 ***	0.791 ***
	(−3.592)	(0.532)	(3.420)	(4.082)
R&D	0.047 **	0.188 ***	0.107 ***	0.077 ***
	(2.400)	(10.213)	(10.681)	(6.470)
FATA	−0.551 **	−0.669 ***	−0.460 ***	−0.405 ***
	(−2.282)	(−2.940)	(−3.560)	(−2.624)
GS	0.044 ***	0.030 ***	0.031 ***	0.034 ***
	(4.066)	(2.966)	(4.860)	(4.388)
HHI	−0.991 **	0.525	0.703 ***	0.779 ***
	(−2.320)	(1.305)	(4.072)	(4.005)
AGE	0.240 ***	−0.094	0.002	0.004
	(2.884)	(−1.192)	(0.046)	(0.071)
Constant	−9.467 ***	−15.360 ***	−17.500 ***	−18.020 ***
	(−13.370)	(−22.135)	(−37.880)	(−25.376)
Year/Ind/Firm FE	Yes	Yes	Yes	Yes
N	2249	2249	2249	2249
R2	0.4053	0.7851	0.4382	0.7402
Sobel Test		0.000 ***		0.004 **

Note: t-statistics are in parentheses; ** p < 0.05, *** p < 0.01.

Table 10. Mechanism test results of downstream supply chain test.

Variable	(1)	(2)	(3)	(4)
	IC_DIS1		IC_DIS2
	Med	IC_CHAIN	Med	IC_CHAIN
IC_FOCAL	−0.076 ***	0.147 ***	0.524 ***	0.810 ***
	(−2.631)	(12.931)	(3.345)	(3.510)
Med	−	−0.017 ***	−	0.236 ***
	−	(−3.082)	−	(5.979)
SIZE	0.193 ***	0.804 ***	0.634 ***	0.801 ***
	(3.796)	(40.348)	(22.033)	(40.576)
LEV	−1.675 ***	−0.125	0.068	−0.053
	(−6.259)	(−1.187)	(0.525)	(−0.510)
ROE	0.153	0.390 *	0.282	0.229
	(0.298)	(1.942)	(1.174)	(1.151)
R&D	−0.068 **	0.074 ***	0.169 ***	0.065 ***
	(−2.141)	(5.977)	(9.013)	(5.275)
FATA	−0.994 **	−0.729 ***	−0.549 **	−0.429 ***
	(−2.424)	(−4.546)	(−2.399)	(−2.674)
GS	0.074 ***	0.039 ***	0.036 ***	0.034 ***
	(3.548)	(4.782)	(3.499)	(4.234)
HHI	−1.505 ***	0.775 ***	0.317	0.808 ***
	(−2.914)	(3.839)	(0.787)	(4.039)
AGE	−0.486 ***	0.017	−0.025	0.102 *
	(−3.616)	(0.314)	(−0.318)	(1.949)
Constant	−0.576	−19.390 ***	−16.190 ***	−19.640 ***
	(−0.306)	(−26.316)	(−24.152)	(−26.887)
Year/Ind/Firm FE	Yes	Yes	Yes	Yes
N	4887	4887	4887	4887
R2	0.1286	0.7466	0.7849	0.7507
Sobel Test		0.032 **		0.001 ***

Note: t-statistics are in parentheses; * p < 0.1, ** p < 0.05, *** p < 0.01.

Table 10. Mechanism test results of downstream supply chain test.

Variable	(1)	(2)	(3)	(4)
	IC_DIS1		IC_DIS2
	Med	IC_CHAIN	Med	IC_CHAIN
IC_FOCAL	−0.076 ***	0.147 ***	0.524 ***	0.810 ***
	(−2.631)	(12.931)	(3.345)	(3.510)
Med	−	−0.017 ***	−	0.236 ***
	−	(−3.082)	−	(5.979)
SIZE	0.193 ***	0.804 ***	0.634 ***	0.801 ***
	(3.796)	(40.348)	(22.033)	(40.576)
LEV	−1.675 ***	−0.125	0.068	−0.053
	(−6.259)	(−1.187)	(0.525)	(−0.510)
ROE	0.153	0.390 *	0.282	0.229
	(0.298)	(1.942)	(1.174)	(1.151)
R&D	−0.068 **	0.074 ***	0.169 ***	0.065 ***
	(−2.141)	(5.977)	(9.013)	(5.275)
FATA	−0.994 **	−0.729 ***	−0.549 **	−0.429 ***
	(−2.424)	(−4.546)	(−2.399)	(−2.674)
GS	0.074 ***	0.039 ***	0.036 ***	0.034 ***
	(3.548)	(4.782)	(3.499)	(4.234)
HHI	−1.505 ***	0.775 ***	0.317	0.808 ***
	(−2.914)	(3.839)	(0.787)	(4.039)
AGE	−0.486 ***	0.017	−0.025	0.102 *
	(−3.616)	(0.314)	(−0.318)	(1.949)
Constant	−0.576	−19.390 ***	−16.190 ***	−19.640 ***
	(−0.306)	(−26.316)	(−24.152)	(−26.887)
Year/Ind/Firm FE	Yes	Yes	Yes	Yes
N	4887	4887	4887	4887
R2	0.1286	0.7466	0.7849	0.7507
Sobel Test		0.032 **		0.001 ***

Note: t-statistics are in parentheses; * p < 0.1, ** p < 0.05, *** p < 0.01.

6. Conclusions

This paper uses the supply-chain portfolio screened from the key industrial chains and focal firms announced by Chinese governments from 2009 to 2024 as the research sample to examine the bidirectional innovation spillovers from focal firms to supply-chain firms. The research results show that there are bidirectional innovation spillovers from focal firms to supply-chain firms. And the innovation spillover effect is asymmetric; that is, compared with the upstream supply chain, the innovation spillover from focal firms to supply-chain firms in the downstream supply chain is stronger. The mechanism test results show that focal firms can enhance the innovation level of upstream firms through the cooperative innovation mechanism, and can also enhance the innovation level of downstream firms through the knowledge spillover mechanism. This study not only enriches the existing literature on the spillover effects of innovation and the role of focal firms in the supply chain, but also provides empirical evidence and theoretical support for focal firms to play a core role, promote innovation in the supply chain, and lead the overall development of the supply chain.

Based on the above research conclusions, this paper draws the following policy implications. First, the government should not only focus on cultivating focal firms and strengthening the supply chain’s collaborative innovation mechanism led by these focal firms, but also establish a dynamic assessment mechanism for focal firms, and regularly adjust the list of focal firms to ensure that they truly possess leading capabilities. Moreover, through tax incentives or subsidies, the government should encourage focal firms to open up innovation resources to the upstream and downstream firms, further enhancing the bidirectional innovation spillovers from focal firms to supply-chain firms. Second, as the core node, focal firms should actively undertake the leading responsibility for supply-chain innovation. By establishing collaborative innovation platforms and optimizing supply-chain finance, they can reduce the innovation costs of supply-chain firms and promote the innovation spillover from focal firms to supply-chain firms. At the same time, considering the asymmetry of innovation spillover effects, focal firms should implement differentiated collaborative innovation strategies for the upstream and downstream supply chains. In the upstream supply chain, they should mainly focus on innovation risks and cooperation coordination, while in the downstream supply chain, they should mainly focus on innovation efficiency and information feedback.