Intelligent Manufacturing and Value Creation in Enterprises: Lessons from a Quasi-Natural Experiment in a Chinese Demonstration Project

Abstract

With the rapid advancement of contemporary information technologies, intelligent manufacturing has emerged as a pivotal direction in the global technological transformation. To empirically examine the impact of intelligent manufacturing on enterprise value creation, this article conducts quasi-natural experiments using Chinese intelligent manufacturing demonstration projects as a sample. Specifically, it focuses on Chinese A-share-listed manufacturing enterprises in Shanghai and Shenzhen from 2011 to 2020. According to the report, the implementation of intelligent manufacturing has a positive influence on enterprise value production. This conclusion remains robust even after undergoing a rigorous testing procedure. Mechanism analysis further reveals that alleviating financial constraints and fostering technological innovation are the two primary avenues through which intelligent manufacturing enhances enterprise value creation. Moreover, the study indicates that regions with favorable business environments experience a more conspicuous boost in enterprise value generation due to intelligent manufacturing. Additionally, businesses in the growth stage are more significantly affected by this phenomenon. Overall, this research not only contributes to the existing body of knowledge on this subject but also offers empirical evidence to support businesses in their endeavors to enhance value creation.

1. Introduction

With the rapid development of modern information technologies such as the Internet, cloud computing, and artificial intelligence, the manufacturing industry is undergoing a significant shift from the digital stage to the intelligent stage. Recognizing this trend, China has explicitly identified intelligent manufacturing as the primary focus for transforming and developing the manufacturing industry through initiatives like the “Made in China 2025” plan. Furthermore, the report from the 20th National Congress of the Communist Party of China emphasized the role of Chinese products in driving high-end development, intelligent upgrading, and the realization of strategies for manufacturing power, network strength, and digital transformation. These policies serve as a roadmap for China’s manufacturing sector to embrace and advance in the realm of intelligent manufacturing in the future.

Intelligent manufacturing, characterized by self-perception, self-learning, self-decision-making, self-execution, and self-adaptation, represents a novel production paradigm that results from the deep integration of modern information and communication technology with advanced manufacturing technology. As a transformative mode of production, the adoption of intelligent manufacturing is poised to yield positive outcomes in terms of enterprise value creation. Despite this, scholars and policymakers have underscored the manifold benefits of intelligent manufacturing on business operations, such as mitigating cost rigidity [1], fostering overall factor productivity growth [2,3] and enhancing business performance [4,5]. However, as with any technological revolution, concerns surrounding technological unemployment have surfaced, constraining some corporate executives from wholeheartedly embracing this transition. While technological advancements may indeed replace certain labor-intensive tasks, it is worth noting that intelligent manufacturing should not be regarded as a panacea for dramatically improving labor productivity. In fact, intelligent manufacturing may exacerbate labor market inequality and exert significant impacts on labor-intensive industries [6]. Conversely, the positive effects of technology upgrades on specific industries may not always be readily apparent. For instance, empirical research examining the computer industry reveals the challenges associated with recuperating the costs incurred by technology upgrades, thereby negating their significant positive impact on enterprise value creation [7]. Consequently, numerous scholars assert the imperative need for comprehensive research on the economic ramifications of intelligent manufacturing [8]. Nonetheless, the existing literature predominantly revolves around investigating the realization pathways of intelligent manufacturing premised on its systemic construction [9] and assessing the level of intelligent macro-manufacturing [10]. Regrettably, financial effects pertaining to intelligent manufacturing have received limited scholarly attention, thereby warranting robust investigation. Consequently, undertaking research to explore the intricate relationship between intelligent manufacturing and enterprise value creation assumes paramount practical significance, as it not only facilitates a comprehensive comprehension of its implementation but also proffers valuable guidance in effecting organizational change and improvement. Additionally, such endeavors contribute to the exploration of factors influencing value creation and present novel perspectives for in-depth analysis aiming to optimize resource allocation.

The contributions of this study are manifold. Firstly, it enriches the existing literature on the economic consequences of intelligent manufacturing by providing empirical evidence at the micro-level, thereby offering a more robust and grounded understanding of the relationship between intelligent manufacturing and enterprise value. In contrast to prior normative and theoretical analyses, this empirical investigation provides direct empirical insights into the link between intelligent manufacturing and value creation. Secondly, by exploring the mechanisms through which intelligent manufacturing impacts enterprise value from the perspectives of financing constraints and innovation investment, this study enhances our understanding of the underlying drivers and pathways through which intelligent manufacturing contributes to value creation. These insights can guide enterprises in overcoming any reluctance and embracing intelligent manufacturing strategies with confidence. Lastly, this study examines the mechanisms and regulatory implications of leveraging intelligent manufacturing for value creation, while also considering the heterogeneity of enterprise value based on marketization levels and life cycles. Such comprehensive analysis enhances our understanding of the dynamics of intelligent manufacturing and its differential impact on enterprise value creation, thereby providing useful insights for both academic researchers and policymakers seeking to further promote intelligent manufacturing initiatives.

In summary, this research provides valuable empirical evidence and insights into the relationship between intelligent manufacturing and enterprise value creation. By offering a nuanced understanding of the mechanisms and heterogeneity associated with this relationship, this study contributes to the scholarly literature on intelligent manufacturing’s economic implications and informs strategic decision making for enterprises and policymakers alike. The remaining portions of the manuscript are organized as follows: The second section provides a comprehensive review and synthesis of the existing literature, along with the presentation of the research hypotheses. The third section delineates the research methodology, encompassing the selection of variables, data sources, and the construction of the econometric model. The fourth section comprises the empirical analysis, entailing parallel trend tests, baseline regression, and robustness checks. The fifth section scrutinizes the mechanisms of action. The sixth section delves into heterogeneous discussions based on the impact mechanisms. Finally, the manuscript concludes with a comprehensive summary and the proposal of pertinent policy recommendations.

2. Theoretical Analysis and Research Hypothesis

Intelligent manufacturing serves as a tangible manifestation of cutting-edge technology, whereby automation and informatization are seamlessly integrated within machines [11]. In contrast to the Internet and information technology, intelligent manufacturing emphasizes the convergence of various technologies and machines across multiple facets of manufacturing, including design, production, management, and service. Its primary focus lies in infusing fresh vitality into enhancing the value creation capacity of the manufacturing industry. This study posits that intelligent manufacturing exerts influence on enterprise value creation by virtue of its impact on financing and innovation.

2.1. Financing Channels

The issue of challenging and costly financing poses a significant barrier to China’s progression towards high-quality economic development [12]. The Schumpeterian innovation theory places significant emphasis on the pivotal role played by the accessibility of funds in facilitating technological innovation [13]. In the midst of an economic downturn, the demand for financing primarily revolves around medium- and long-term funds which come with high-risk premiums and various associated expenses [14]. This exerts tremendous pressure on the sustainable development and value creation of enterprises, particularly within the manufacturing industry. Fundamentally, the difficulty in accessing financing stems from traditional financial institutions’ focus on the scale and prospects of enterprises [15]. Intelligent manufacturing, with its promising prospects, holds significant importance on China’s path towards becoming a global scientific and technological powerhouse. Without a doubt, the government will enact a series of policies aimed at supporting intelligent manufacturing. By implementing these policies, intelligent manufacturing enterprises can benefit from government subsidies. Consequently, on the one hand, these enterprises can directly inject research and development funds into their ventures, thereby enhancing the efficiency of resource allocation for innovation and alleviating financial constraints on businesses in accordance with the resource-based theory. On the other hand, by receiving government subsidies, intelligent manufacturing enterprises can send a positive signal of “official certification” to the market, in line with the signal theory. This attracts external financing and stimulates increased investment from financial institutions into these enterprises [16]. Furthermore, intelligent manufacturing offers the potential to optimize both upstream and downstream relationships within the supply chain, consequently mitigating financing constraints. Within the intelligent manufacturing framework, enterprises can leverage the capabilities of emerging information technologies to acquire market data pertaining to the supply chain and potential customers. This allows for expanding strategic partnership networks and utilizing digital platform operation mechanisms, thereby enhancing the visualization of supply chain information. Such advancements enable enterprises to better cater to the needs of both internal stakeholders and customers along the supply chain [17]. The enhanced relationships within the supply chain consequently result in stable and diverse sources of capital inflow, thus alleviating the financing constraints encountered by enterprises in the process of intelligent transformation and upgrade.

2.2. Innovation Channels

The driving force behind the innovation capability of intelligent manufacturing enterprises lies in informatization and intelligentization [18]. Firstly, intelligent manufacturing effectively addresses issues of information asymmetry and communication by leveraging advanced intelligent equipment and various artificial intelligence technologies. Intelligent manufacturing plays a crucial role in breaking down information barriers across different departments within a company, thus facilitating seamless information exchange. Two powerful tools in achieving this are the Internet of Things (IoT) and cloud services. The IoT enables seamless connectivity and communication between physical devices and networks, while cloud services provide the capability to remotely store real-time operational data, allowing for the complete sharing of essential information needed for production and manufacturing processes [19]. Furthermore, leveraging cloud computing and data analysis technologies within the framework of intelligent manufacturing allows for the comprehensive processing and analysis of information, effectively aligning it with the information processing needs brought upon by the company’s external environment [20,21,22].

This leads to enhanced synchronicity between product innovation and market innovation, thereby boosting enterprise value [23]. Additionally, intelligent manufacturing facilitates intelligent real-time coordination and the sharing of information resources across all production links within enterprises. Furthermore, its efficient information searching and processing capabilities enable the automation of mundane and repetitive tasks, allowing researchers to dedicate more time to complex and precise R&D and innovation activities [24]. This fosters the effectiveness of enterprise innovation. Moreover, intelligent manufacturing revolutionizes resource allocation methods and enhances resource conversion efficiency [25]. By enabling rapid production of new products through the replacement of a few parts or components, it significantly diminishes switching costs and frees up more funds for investment in innovative projects, thereby bolstering long-term enterprise value. Simultaneously, intelligent manufacturing liberates workers from mundane tasks and reduces labor costs (10), allowing for increased investment in complex and sophisticated R&D endeavors. This, in turn, contributes to the overall enhancement of enterprise value.

In summary, intelligent manufacturing exerts its influence on enterprise value creation by mitigating financing constraints and promoting enterprise innovation.

The theoretical framework representing this relationship is illustrated in Figure 1.

Building upon this basis, we propose the central hypothesis for this study as follows:

Hypothesis 1.

Intelligent manufacturing will foster the creation of enterprise value.

Figure 1. The logical framework of intelligent manufacturing promoting enterprise value creation.

Figure 1. The logical framework of intelligent manufacturing promoting enterprise value creation.

3. Research Design

3.1. Sample Setting and Data Source

3.1.1. Research Methods

The theoretical framework of the difference-in-differences (DID) method is built upon the foundation of “natural experiments”. The concept of natural experiments commonly used in the field of economics in our country is often referred to as quasi-experiments. These two terms have subtle yet significant distinctions—natural experiments are completely random “natural events”, while quasi-experiments involve the deliberate selection and grouping of experimental samples by the researchers themselves [26]. Due to the inability to select and group experimental subjects completely at random within an experimental environment, quasi-experiments strive to approximate randomness in sample selection and grouping.

From 2015 to 2018, in order to accumulate practical experience in intelligent manufacturing implementation, the Ministry of Industry and Information Technology (MIIT) of our country selected a group of enterprises each year as pilot units for intelligent manufacturing projects. The selection of enterprises as intelligent manufacturing pilots by the MIIT was conducted gradually on an annual basis. As an exogenous event, the policy of intelligent manufacturing pilot projects can be approximated as a random event, meaning that the sample selection for this quasi-experiment is close to random, exhibiting the characteristics of a quasi-natural experiment. The MIIT’s selection of pilot enterprises not only provides objective variables for measuring intelligent manufacturing but also creates an opportunity for a quasi-natural experiment that can overcome endogeneity through the difference-in-differences method (DID).

The difference-in-differences (DID) method effectively addresses endogeneity issues caused by omitted variables but requires exogenous shocks to be random. However, in this study, the MIIT may require candidate enterprises for pilot selection to have already demonstrated initial achievements in intelligent development and maintain good business performance. This may lead to the pilot enterprises themselves having a higher value of development compared to the control group enterprises. Therefore, there may be a self-selection issue in the selection of intelligent manufacturing pilots. The propensity score matching (PSM) method proposed by Heckman et al. (1997) [27] can identify samples that are as similar as possible to the treatment group samples in terms of individual characteristics, providing a countermeasure to address sample selection bias. Specifically, based on the individual characteristics of the samples, this study matches the control group and treatment group samples and then removes samples that are not in the common support region, thereby constructing an experimental group and a control group with basic homogeneity of individual characteristics, except for differences in the implementation of intelligent manufacturing. This ensures more reliable results of the difference-in-differences analysis.

3.1.2. Sample Selection and Data Source

In accordance with the guidelines provided in the Intelligent Manufacturing Engineering Implementation Guide (2016–2020) and other relevant documents, the Ministry of Industry and Information Technology selects a designated number of enterprises annually, from 2015 to 2018, based on their compliance with the criteria outlined in the Smart Manufacturing Demonstration Project. This research aims to examine the impact of the brilliant manufacturing pilot initiative by selecting Chinese A-share-listed manufacturing companies as the research sample, covering the period from 2011 to 2020. The pilot initiative is considered a quasi-natural experiment, wherein a manual compilation of manufacturing firms that have implemented smart manufacturing policies is conducted. These firms constitute the treatment group, while the remaining firms form the control group.

To gather the fundamental characteristics and financial information of the selected manufacturing listed companies, data from the Guotaian CSMAR and Wind databases are utilized. Following standard research protocols, the data screening and processing procedures are executed as follows: (1) Excluding stocks of domestic publicly listed companies that have incurred financial losses for three consecutive years and have received risk warnings regarding potential delisting, as well as those with incomplete statistical year data. (2) The dataset is refined to include 16,487 firm-annual observations after eliminating missing data and outliers. Additionally, continuous variables are Winsorized at the 1% and 99% quantiles. Among these observations, 2420 firms belong to the control group (representing firms that did not adopt innovative manufacturing strategies, assigned a value of 0), while 81 firms belong to the treatment group (representing firms that adopted innovative manufacturing strategies, assigned a value of 1).

3.1.3. Sample Matching

To address the issue of sample selection bias, this study applies the propensity score matching (PSM) method proposed by Heckman et al. [27] to process the primary data. The following steps are taken:

(1)

Matching Method:

Variables that are common between the treatment and control groups are selected for matching purposes. Additionally, samples that do not fall within the support domain of the control group but possess the most similar matching characteristics to the treatment group are also considered for PSM. This approach draws explicit reference from the work of Tang Song et al. [28]. The propensity score is estimated using logit regression, and the samples returned are compared to identify the nearest 1:1 neighbor for screening and matching.

(2)

Matching Results:

Table 1. Results of the balance test.

Variables	Treatment	Means		SE (%)	Reduction of Standard Deviation (%)	t-Test
		Treatment Group	Control Group			t	p > |t|
ROA	U	0.0576	0.0523	8.000	76.6	1.370	0.172
	M	0.0576	0.0563	1.900		0.230	0.821
Size	U	23.27	21.95	107.4	98.90	20.15	0.000
	M	23.27	23.28	−1.200		−0.140	0.891
Age	U	19.87	17.07	53.70	92.50	8.950	0.000
	M	19.87	20.08	−4.000		−0.500	0.616
Lev	U	0.452	0.383	36.80	87.20	6.210	0.000
	M	0.452	0.461	−4.700		−0.580	0.563
growth	U	0.180	0.199	−2.500	16.20	−0.370	0.712
	M	0.180	0.163	2.100		0.340	0.734
Dual	U	1.674	1.663	2.200	−131.7	0.390	0.694
	M	1.674	1.700	−5.000		−0.620	0.534
Mshare	U	0.0416	0.0921	−38.40	93.10	−5.750	0.000
	M	0.0416	0.0451	−2.600		−0.410	0.680

presents the results of the matching process, showing that the absolute deviation of variables after matching is less than 9%, where U represents the unmatched result and M represents the matched comparison result. This indicates a high level of comparability between the samples before and after matching. Furthermore, after matching, the t-test conducted on each characteristic variable no longer yields significant results, confirming the validity of the matching outcome. Figure 2 and Figure 3 illustrate the probability distribution between the treatment group and the control group after matching, revealing a substantial increase compared to the distribution before matching. This provides further evidence of the effectiveness of the matching results.

Table 2. Definition and measurement of variables.

Variables	Signs	Description of Variable Valuation Method
Enterprise value	EVA	Net operating profit after tax—adjusted capital × Weighted average cost of capital
Intelligent manufacturing	DID	Suppose the enterprise is selected as the demonstration unit of intelligent manufacturing for the first time in the year. In that case, the value of DID will be 1 in the current year and subsequent years; otherwise, 0.
Return on assets	ROA	Net profit/Average balance of total assets
Company size	Size	Natural logarithm of total assets
Enterprise age	Age	Years of listing
Debt ratio	Lev	Total liabilities/Total assets
Enterprise growth	growth	The annual growth rate of operating income
duality	Dual	The Chairman and General Manager shall be the same person with a value of 1, otherwise 0
Executive shareholding	Mshare	Number of senior management holdings/total shares

presents the definition and measurement of variables.

3.2. Model Construction and Variable Definition

This study explores the impact of intelligent manufacturing on firm value generation by employing an undifferentiated treatment group and a control group created through propensity score matching. To assess the accuracy of hypothesis, a model (1) is constructed based on the research conducted by Anderson et al. [29].

$${EVA}_{i,t}=\alpha +\beta {DID}_{i,t}+\gamma {Controls}_{i,t}+{\delta }_i+{\epsilon }_t+{ϵ}_{i,t}$$

(1)

In this context, “i” represents the firm, while “t” denotes the year. EVA refers to the level of value creation within the firm. DID is a binary variable indicating whether the firm has implemented smart manufacturing policies. Specifically, it takes on a value of 1 if the firm is selected as a smart manufacturing demonstration unit in the current year or subsequent years; otherwise, it is coded as 0. The term “Controls” refers to the control variables employed in this study. Additionally, $\delta$ and $\epsilon$ are employed to capture both firm-specific and year-specific fixed effects, $ϵ$ reflecting random disturbances. Robust standard errors are utilized in the empirical analysis to address any potential heteroscedasticity concerns. The primary focus of this paper lies in the coefficient $\beta$ of the main explanatory variable, DID. If $\beta$ is found to be statistically significant and positive, it would provide evidence supporting the hypothesis that smart manufacturing significantly fosters firm value (Hypothesis).

4. Empirical Results and Analysis

4.1. Descriptive Statistics

Differences can be observed in the levels of development across various enterprise values, as evidenced by the range of enterprise values presented in

Table 3. Descriptive statistics.

Variable	Sample	Mean	SD	Median	Maximum	Minimum
EVA	16,311	1.0332	6.0034	0.1099	39.3329	−15.6097
DID	16,311	0.0192	0.1372	0.0000	1.0000	0.0000
ROA	16,311	0.0534	0.0646	0.0500	0.2330	−0.2025
Lev	16,311	0.3831	0.1928	0.3710	0.8571	0.0508
Dual	16,311	1.6733	0.4690	2.0000	2.0000	1.0000
Ind	16,311	0.3750	0.0531	0.3333	0.5714	0.3333
Size	16,311	21.9838	1.1395	21.8177	25.4287	20.0527
Age	16,311	17.1585	5.4468	17.0000	31.0000	5.0000
Audittyp	16,311	0.0227	0.1489	0.0000	1.0000	0.0000

, spanning from 39.332 to −15.609. The average DID score of 0.0192 indicates that approximately 1.92% of the sampled businesses have embraced intelligent manufacturing. This discovery underscores the existence of ample room for enhancing the implementation of intelligent manufacturing policies within manufacturing enterprises. Furthermore, the distribution of other control variables falls within a reasonable range.

4.2. Benchmark Regression

Table 4. Benchmark regression.

	Full Sample		PSM
Variable	(1)	(2)	(3)	(4)
DID	4.1726 **	4.0635 **	2.8501 ***	2.8200 ***
	(2.4390)	(2.5448)	(3.0737)	(3.4513)
ROA		56.3227 ***		44.7516 ***
		(12.5885)		(25.5997)
Size		1.1896 *		0.8095 ***
		(1.8416)		(3.9376)
Age		0.3738		0.2349 *
		(1.5960)		(1.9579)
Lev		−0.7169		0.3580
		(−0.5571)		(0.6677)
Growth		−0.2667 **		−0.0701
		(−2.2701)		(−1.0726)
Dual		−0.5318		−0.0346
		(−1.5272)		(−0.2435)
Mshare		−0.6358		0.6233
		(−0.6897)		(1.4359)
_cons	0.5435 **	−32.0288 **	0.3428 ***	−23.1857 ***
	(2.4113)	(−2.4106)	(2.6079)	(−4.9259)
N	16,487	16,487	16,311	16,311
R2	0.0176	0.1217	0.0424	0.3028
Adj. R2	0.0170	0.1208	0.0419	0.3021

Note: ***, **, and * denote a 1%, 5%, and 10% level of significance, respectively.

presents the double difference estimation using both the full sample and the sample matched through propensity score matching (PSM). In the first column, the regression results are shown without controlling for covariates in the full sample mode, allowing for a direct comparison with the second column, which includes the covariates in the full sample mode for a more comprehensive analysis. The third and fourth columns compare the results before and after matching the samples using PSM, while also incorporating the covariates.

Upon comparing the two models without controlling for covariates, it is evident that the regression coefficients of the pilot policy in the model with covariates are consistently lower. This finding provides empirical support for the reasonableness of selecting and including covariates in the analysis. The results demonstrate a significant positive relationship between the adoption of an intelligent manufacturing strategy and the generation of enterprise value, regardless of the inclusion of control variables. This implies that companies implementing intelligent manufacturing practices are more effective in promoting corporate value growth compared to those that do not embrace such practices. These findings substantiate the initial theoretical proposition; intelligent manufacturing is a responsive system that utilizes information technology to achieve timely and efficient implementation at the shop floor and higher levels. It leverages advanced information and manufacturing technologies to enhance the flexibility of manufacturing processes and strengthen market demand forecasting [30]. As a result, businesses can benefit from reduced financing costs, increased innovation in technology, and ultimately achieve higher value creation. Intelligent manufacturing is essentially the result of deep integration between the Internet and the manufacturing industry. The reason why intelligent manufacturing technologies can enhance the value creation level of manufacturing enterprises is not only due to their ability to reduce financing costs, but also because industrial Internet improves the flexibility of production equipment, greatly increasing the variety of manufacturing products and making personalized manufacturing possible. Manufacturing enterprises not only benefit from economies of scale achieved through specialization but also experience bidirectional growth in economies of scope [31]. In the digital economy era, the increasing differentiation, personalization, and customization preferences in the demand side of society are driving enterprises to enhance flexible production or intelligent manufacturing technologies based on big data. Promoting the application of artificial intelligence in the physical economy sector encourages enterprises to comprehensively utilize techniques such as intelligent production decision making and flexible manufacturing to optimize resource allocation and input–output efficiency during the production process. By doing so, enterprises can enhance value creation and facilitate the widespread application of intelligent manufacturing models through pilot demonstrations.

Moreover, the regression coefficients of other control variables align with prior research findings. Specifically, the ROA (return on assets) and size coefficients exhibit a significantly positive association with value creation, implying that these variables contribute significantly to enhancing the value generation of businesses. Conversely, variables such as dual, age, and Mshare do not show any discernible effect on enterprise value creation, indicating that they are not the primary factors influencing the rise in value. Overall, these empirical findings provide valuable insights into the relationship between intelligent manufacturing, business value creation, and the influence of control variables, contributing to the existing literature on this subject.

4.3. Robustness Test

4.3.1. Balance Trend Test

The objective of utilizing the double difference model analysis in this study is to assess the validity of the “parallel trend” hypothesis. This hypothesis suggests that prior to implementing an intelligent manufacturing policy, the treatment group and control group should exhibit a parallel growth trajectory in terms of enterprise value creation. In order to examine whether this hypothesis holds true, the paper draws inspiration from the event study method employed by Beck et al. [32] and constructs Model 2. The purpose of this model is to evaluate whether a statistically significant distinction between the pre- and post-implementation periods of the intelligent manufacturing policy group and the control group exists, thus enabling an accurate determination of the extent to which enterprise value creation adheres to the parallel trend hypothesis.

$${EVA}_{i,t}=\alpha +{\sum }_{n=1}^7{\beta }_{k}Before{N}_{it}+\theta {After}_{it}+\gamma {Controls}_{it}+{\delta }_i+{\epsilon }_t+{ϵ}_{i,t}$$

(2)

This study refers to the research conducted by Beck et al. [32] to construct a parallel trends test model, as depicted in Equation (2). In this model, the variable “Before1” takes a value of 1 to indicate that the treatment group samples were collected in the year prior to the pilot selection, and 0 otherwise. The variables “Before2” through “Before7” follow the same logic. The variable “After” takes a value of 1 to represent the treatment group samples from the year of the pilot selection and onwards, and 0 otherwise. The definitions of other variables remain consistent with those mentioned earlier. Following the methodology used by Beck et al. [32], the year of pilot selection is considered the base year. The parameter ${\beta }_k$ reflects the size of the difference in firm value between the treatment and control groups before the pilot selection. The results of the parallel trends test are illustrated in Figure 4.

4.3.2. Assessing the Efficacy: A Placebo Test Analysis

To further investigate the influence of unobservable factors on enterprise value creation, a placebo test was conducted in a randomized experimental group. A total of 500 random samples were selected from each enterprise’s corresponding year variable, representing the policy impact point, and re-estimated using model (1). Figure 5 illustrates the distribution of regression estimation coefficients and post-randomization p values. The majority of estimations demonstrate p values exceeding 0.1, indicating that the mean estimation coefficient values tend to cluster around zero. This compelling observation reinforces the results of the placebo test, suggesting that the positive effect on enterprise value generation through piloting an intelligent manufacturing program is not coincidental.

4.3.3. Testing the Resilience: Exploring Additional Robustness Measures

The present study includes several robustness tests to enhance the validity of the findings. These tests are designed to address potential confounding factors and verify the consistency and robustness of the results. The tests are as follows:

(1) Replacement of interpreted variables: Building on the research conducted by Liu Xing and Li Xiaorong [33], this study employs the Tobinq value as an alternative measurement index for enterprise value creation in regression analysis. This substitution yields consistent regression results that align with the previously discussed findings.

(2) Modification of matching methodology: In this paper, the matching method used in the initial analysis is replaced with a proximity-matching approach to conduct core matching on the original research samples. The outcomes, as illustrated in

Table 5. Results of the robustness test.

Variable	Tobinq	Kernel Matching	Control Industry Year
	(1)	(2)	(3)
DID	0.2800 ***	4.0635 **	2.8501 ***
	(2.8612)	(2.5448)	(3.0737)
Controls	Y	Y	Y
_cons	18.8511 ***	−32.0288 **	0.3428 ***
	(16.6669)	(−2.4106)	(2.6079)
Year	Y	Y	Y
Firm	Y	Y	Y
Year×Ind	N	N	Y
N	16,311	16,487	16,311
R2	0.3940	0.1217	0.0424
Adj. R2	0.3934	0.1208	0.0419

Note: *** and ** denote a 1% and 5% level of significance, respectively.

, column (2), affirm that intelligent manufacturing continues to significantly support the value development of manufacturing businesses, even after modifying the core matching mechanism. This finding is consistent with the results obtained from the application of double difference estimation after adjacent matching, thereby reinforcing the robustness of the conclusions.

(3) Control for macroeconomic policy and industry characteristics: To mitigate potential confounding effects stemming from macroeconomic policy implementation and cyclical changes in the manufacturing industry, this study adopts methods proposed by Pan Yue [34] and others to control for high-level combined fixed effects of the year. The test results, presented in Table 5 (3), convincingly demonstrate that the positive impact of the difference-in-differences (DID) estimate on enterprise value creation persists at a significant level of 1%, even after accounting for external economic policies. This finding underscores the robustness of the conclusions while considering factors such as macroeconomic policies and industry characteristics.

Collectively, these robustness tests provide supplementary evidence to support the findings of the study and bolster the validity of the derived conclusions.

5. Test of Action Mechanism

Based on the aforementioned analysis results, this paper aims to further explore the role of intelligent manufacturing in the creation of enterprise value. Theoretical analysis suggests that intelligent manufacturing influences value creation through its impact on financing constraints and technological innovation. To examine this, the research methodology employed in Jiangtao’s study is adopted [35], consisting of two steps: (1) Assessing the influence of intelligent manufacturing on financing constraints and enterprise technological innovation. Supported initially by the theoretical logic, intelligent manufacturing is expected to alleviate financing constraints and enhance the quality of innovation within enterprises. (2) Building upon the existing literature that highlights the direct and evident effects of financing constraints and technological innovation on enterprise value creation, this paper further investigates how intelligent manufacturing specifically contributes to the overall process.

5.1. Financing Channels

From a theoretical perspective, intelligent manufacturing has a significant impact on alleviating financing constraints within enterprises (S.A.). Firstly, intelligent manufacturing strengthens the enterprises’ ability to receive government subsidies, thereby enhancing their access to external financing. Secondly, intelligent manufacturing facilitates the development of high-quality financial services tailored to manufacturing businesses, such as supply chain finance, which promotes the integration of the economic value chain. To further validate this mechanism, this paper utilizes the observed enterprise S.A. index to measure financing constraints, applying the methods proposed by Hadlock and Piere [36]. $S.A.=-0.737\times Size+0.043\times {Size}^2-0.04\times Age$. Empirical results are presented in column (1) of

Table 6. Mechanism analysis.

	S.A.	R.D.
Variables	(1)	(2)
DID	−0.0352 ***	0.0587 **
	(−4.8269)	(1.9764)
Controls	Y	Y
_cons	2.6754 ***	1.8640 ***
	(27.4425)	(3.7730)
Year	Y	Y
Firm	Y	Y
N	16342	15064
R2	0.9079	0.0755
Adj. R2	0.9078	0.0746

Note: *** and ** denote a 1% and 5% level of significance, respectively.

, illustrating the impact of intelligent manufacturing on financing constraints in enterprises. The results reveal a significantly negative coefficient for the DID variable at a 1% confidence level, confirming the substantial alleviation of financing constraints through intelligent manufacturing. Furthermore, existing research supports the positive effect of mitigating financing constraints on enterprise value development. The incomplete financial market has long plagued enterprises with the high costs of external financing and hindered them from achieving optimal investment portfolios [37]. Severe financial constraints often lead to underinvestment by businesses, prompting inappropriate and inefficient investment decisions that impede the growth of enterprise value. In this context, intelligent manufacturing serves as a catalyst for enterprise value creation by alleviating financing constraints, offering diverse opportunities for business growth.

5.2. Innovation Channels

Enterprise technological innovation is characterized by high investment costs, a long return period, and significant uncertainty [38]. These factors often hinder businesses from accurately predicting future profits, resulting in limited investment in innovation. However, intelligent manufacturing policies have the potential to enhance an enterprise’s capabilities in information gathering and processing, thus reducing the uncertainty associated with the innovation process and increasing the effectiveness of innovation. Additionally, intelligent manufacturing enables the transformation of labor cost savings from repetitive tasks into investments in high-skilled talent, which plays a critical role in facilitating enterprise innovation amidst rapidly evolving technology.

To empirically test the impact of intelligent manufacturing on technological innovation, this paper utilizes the proportion of R&D expenditure to total assets as a measure of the enterprise’s level of technological innovation (R.D.). The results presented in column (2) of Table 6 demonstrate a significantly positive coefficient for the DID variable at a 5% level of significance. This finding indicates that intelligent manufacturing has a significant positive effect on the technological innovation of enterprises. These empirical findings align with the previous theoretical analysis that highlights the positive impact of intelligent manufacturing on technological innovation within enterprises. Furthermore, existing research provides robust academic support for the relationship between technological innovation and enterprise value creation [39]. According to Schumpeter’s innovation theory, technological innovation is instrumental in driving sustainable enterprise development. The decision-making process of innovation within enterprises is closely intertwined with the development of enterprise value [40]. Furthermore, manufacturing enterprises can enhance their enterprise value by expanding market share through R&D innovation and differentiation strategies [41].

6. Further Analysis

This paper contributes to the research on the impact of intelligent manufacturing on enterprise value creation by conducting a comprehensive analysis from a heterogeneity perspective, considering the degree of market-oriented development and enterprise characteristics. By doing so, it aims to provide new empirical evidence for a more accurate understanding of the value creation enabled by intelligent manufacturing.

6.1. Business Environment

Given the evident variations in the degree of the business environment across different provinces in China, as well as the diverse external market and legal environments faced by enterprises in different regions, it is crucial to further analyze the disparities in the implementation of intelligent manufacturing policies within different business environments. Theoretically, the availability of a supportive environment and a technological innovation base for manufacturing enterprises, along with the self-interest motivation of enterprise management to pursue intelligent transformation and upgrading, are factors that may not be uniformly prevalent. Therefore, the impact of smart manufacturing on enterprise value development is anticipated to be more pronounced in regions with relatively underdeveloped business environments.

To support this proposition, this research refers to the China Sub-Provincial Marketization Index Report by Wang Xiaolu et al. [42], which provides an analysis of the local business environment (market) and categorizes the sample into two groups based on the median. Referring to

Table 7. Further analysis.

	Market-Oriented Process		Life Cycle
Variable	High	Low	Growth Period	Non-Growth Period
DID	2.1266	3.2935 ***	3.1407 ***	−1.6228 **
	(1.6225)	(3.0045)	(3.5360)	(−2.0558)
Controls	Y	Y	Y	Y
_cons	−28.5334 ***	−17.6341 **	−23.5835 ***	−15.2287
Year	Y	Y	Y	Y
Firm	Y	Y	Y	Y
N	8538	7773	14,547	1764
R2	0.2949	0.3158	0.3081	0.2792
Adj. R2	0.2935	0.3143	0.3073	0.2722

Note: *** and ** denote a 1% and 5% level of significance, respectively.

, columns (1) and (2), the study finds that the promotion of smart manufacturing is more significant in regions characterized by a slower pace of business environment development. This finding suggests that the implementation of smart manufacturing strategies could be a viable approach to further advancing firm development in such regions.

In summary, this research recognizes the importance of considering the heterogeneity of business environments and argues that the impact of intelligent manufacturing on enterprise value creation varies across different regions. By drawing on empirical evidence from the China Sub-Provincial Marketization Index Report, this research supports the notion that intelligent manufacturing promotion is more prominent in regions with a relatively less-developed business environment. These findings highlight the potential benefits of implementing smart manufacturing strategies to foster further firm growth and development.

6.2. Enterprise Characteristics

Enterprises serve as the key implementers of intelligent manufacturing policies. By further examining the cross-sectional differences in supporting enterprise value creation through intelligent digital manufacturing based on enterprise characteristics, we can ascertain the accuracy of our conclusions and provide policy recommendations. In line with the enterprise life cycle theory, businesses evolve through distinct stages, including start-up, growth, maturity, and decline. At different stages of their life cycles, businesses face varying financial constraints and engage in different levels of technological innovation, which in turn influence the effectiveness of intelligent manufacturing in generating value.

Building on the insights of Li Tang and other studies [43], this research adopts a definition where a growth enterprise signifies a duration of more than ten years and assigns a value of 1 or 0 accordingly. The focus is primarily on the heterogeneity of intelligent manufacturing between the growth and non-growth periods of enterprises. As illustrated in columns (3) and (4) of Table 7, during the growth stage, companies primarily emphasize expanding their operational scale and capturing market share, resulting in a substantial demand for intelligent transformation. The impact of intelligent manufacturing on the growth of the enterprise value is significantly higher (coefficient 3.1407, with a 1% significance level) for growth firms compared to non-growth firms (coefficient −1.6228, with a 5% significance level), showcasing clear disparities between these two groups.

The results highlight that enterprises in the growth stage are more likely to take the lead in embracing the wave of intelligent transformation after accumulating technology and capital during the early stages. Conversely, young non-growing enterprises have yet to complete the initial capital accumulation phase and may possess relatively limited resources. Hence, they should exercise caution in selecting intelligent manufacturing policies to maximize value creation. In conclusion, this research emphasizes the significance of considering the heterogeneity of enterprise characteristics when examining the relationship between intelligent manufacturing and enterprise value creation. The findings support the notion that the impact of intelligent manufacturing differs based on the growth stage of enterprises. This knowledge provides valuable insights for guiding policy recommendations tailored to meet the specific needs and challenges encountered by different types of enterprises.

7. Research Conclusions and Policy Recommendations

To investigate and analyze the mechanism and impact of intelligent manufacturing on corporate value generation, this study adopts a quasi-natural experiment approach utilizing the pilot demonstration project of intelligent manufacturing. By analyzing data from Chinese manufacturing listed companies spanning the period from 2011 to 2020, the study provides empirical evidence that the deployment of intelligent manufacturing significantly enhances long-term enterprise value creation. This conclusion remains robust even after subjecting it to rigorous testing through parallel trend analysis and placebo tests, ensuring the reliability of the findings.

Furthermore, the study conducts a thorough mechanism analysis, revealing that financing constraints and technological innovation serve as key channels through which intelligent manufacturing promotes enterprise value creation. Additionally, the research delves into the nuanced impact of intelligent manufacturing policy on value generation, taking into account the stage of the marketization process and the unique attributes of businesses. The results demonstrate that the promotion of an intelligent manufacturing strategy for value creation is particularly effective during the growth stage of companies in a gradual marketization process.

In summary, this study contributes valuable insights into the mechanisms and impacts of intelligent manufacturing on corporate value generation. By leveraging a quasi-natural experiment and utilizing data from Chinese manufacturing listed companies, the study confirms that intelligent manufacturing significantly promotes long-term enterprise value creation. Furthermore, the study highlights the importance of addressing financing constraints and fostering technological innovation as key drivers of value creation. Finally, it underscores the need for tailored intelligent manufacturing policies that align with the specific growth stages and marketization processes of businesses to maximize value generation potential. Practitioners in the field can derive several benefits from the research on intelligent manufacturing presented in this study. Firstly, the study provides empirical evidence that the deployment of intelligent manufacturing significantly enhances long-term enterprise value creation. This finding implies that practitioners who adopt intelligent manufacturing practices can expect to improve their company’s overall value and long-term financial performance. By incorporating advanced technologies and automation into their manufacturing processes, practitioners can streamline operations, increase efficiency, and ultimately enhance their competitiveness in the market. Secondly, the mechanism analysis conducted in the study reveals that financing constraints and technological innovation are key channels through which intelligent manufacturing promotes enterprise value creation. This insight is valuable for practitioners as it emphasizes the importance of addressing financing challenges and investing in technological advancements. By recognizing these factors, practitioners can strategically allocate resources and make informed decisions to overcome financial limitations and drive innovation within their organizations. Moreover, the study highlights the nuanced impact of intelligent manufacturing policies on value generation. It highlights the significance of considering the stage of the marketization process and the unique attributes of businesses when implementing intelligent manufacturing strategies. This finding provides practitioners with a tailored approach to maximize the benefits of intelligent manufacturing based on their specific growth stages and market conditions. Overall, practitioners can benefit from this research by gaining a deeper understanding of the mechanisms and impacts of intelligent manufacturing on corporate value generation. By leveraging the insights and empirical evidence provided, practitioners can make informed decisions, allocate resources effectively, address financing constraints, foster technological innovation, and develop customized intelligent manufacturing strategies that align with their specific business contexts. These actions can ultimately lead to improved long-term enterprise value and enhanced competitiveness in the market.

The study derives the following policy suggestions from the aforementioned conclusions: (1) The adoption of intelligent manufacturing will significantly impact the growth of manufacturing businesses’ value. Therefore, it is crucial for the current administration to prioritize the intelligent manufacturing policy, aiming to guide the intelligent transformation of companies. This can be achieved by promoting multi-level and systematic support policies that facilitate the intelligent transformation of enterprises. Additionally, establishing an environment conducive to the intelligent transformation of businesses from a policy and regulatory perspective is essential. (2) Enterprises should recognize the value growth potential associated with the promotion of intelligent transformation. It is important for businesses to leverage their own initiative and prioritize improving financing efficiency and innovation levels. To achieve long-term value development, enterprises should actively cooperate with the national intelligent manufacturing strategic layout. They should integrate intelligent manufacturing into all production processes based on their operational foundations and development requirements. (3) Effective implementation of the intelligent manufacturing policy should thoroughly consider the influence of external environmental factors. Intelligent manufacturing should be seen as a powerful driver for businesses in weak market environments and during high-quality development in the initial growth stage. Businesses in the growth stage, as well as markets with slower development, should fully grasp the advantages of promoting intelligent manufacturing transformation. Early planning, proactive exploration, and the prompt realization of intelligent transformation and upgrading are paramount for attaining high-quality business growth.

In summary, this report provides policy suggestions based on the conclusions drawn. These suggestions emphasize the importance of prioritizing the intelligent manufacturing policy, fostering enterprise-level initiatives, and aligning with the national strategic vision. Additionally, the suggestions highlight the need to consider external environmental factors and the specific stage of business growth to maximize the benefits of intelligent manufacturing transformation for long-term value development.