How Digital Transformation Affect Green Innovation Performance of MNEs: From the Organizational Learning Perspective

Abstract

The digital transformation of multinational enterprises (MNEs) in emerging economies has gained increasing interest in academic circles and business communities. However, how and when digital transformation affects green innovation in MNEs remains unknown. Using the theory of organizational learning, this study explores the relationship between digital transformation and the green innovation performance of MNEs, reveals the underlying theoretical mechanisms (i.e., absorptive capacity), and identifies the moderating role of the degree of internationalization and state ownership. Data from Chinese listed MNEs from 2010 to 2019 were used to examine mediating and moderating effects, and robustness tests were performed. The results show that (1) digital transformation can enhance green innovation in MNEs by improving their absorptive capacity; (2) a high degree of internationalization can strengthen the positive relationship between digital transformation and absorptive capacity; and (3) the positive effect of digital transformation on absorptive capacity is stronger in state-owned MNEs than in private MNEs. This study provides a comprehensive theoretical framework for understanding how and when digital transformation affects green innovation in MNEs. The results provide insights into digital transformation and can inform policies regarding green innovation performance and sustainable development in firms.

1. Introduction

In the age of globalization, green innovation is becoming increasingly vital for multinational enterprises (MNEs), especially those that are striving to overcome technological barriers and obtain a competitive edge in the global marketplace. Meanwhile, facing unprecedented global environmental challenges, firms are increasingly recognizing that traditional approaches to environmental protection are inadequate to meet modern sustainability standards, especially for MNEs, which face dual pressures—strict environmental regulations and expectations from the host countries, alongside growing demands from stakeholders in their host and home countries [1]. Green innovation has emerged as a vital pathway for reducing resource consumption and minimizing ecological damage [2,3]. However, MNEs in emerging markets are lagging behind those in developed economies in terms of technology and strategic management [4]. Most of these MNEs are at a disadvantage in key fields of green innovation, intensifying the need for innovative breakthroughs in emerging economies [5]. The emergence of “Blockchains” and the “Internet of Things” has created a new technological climate for firms [6]. In light of this trend, it is crucial to investigate the effects of digital transformation on green innovation performance and the mediating mechanism by which digital transformation operates.

Previous studies on green innovation performance mainly centered on the motivations [7] and operational mechanisms of MNEs in conducting green innovation [8,9,10,11,12]. Previous research classified motivations into market-driven and resource-based motivations. The main rationale behind market-driven motivation is that MNEs can expand to overseas markets by following consumer trends, engaging with overseas R&D institutions, and developing new products for the host markets [7]. The primary rationale behind resource-based motivation is that, through internationalization, MNEs can acquire resources from the global market for their own green innovation activities [13]. Studies in this area mainly explore factors that promote or hinder green innovation in MNEs. For example, macro-level research focuses on factors such as institutions, cultures, and governments [8,9,14], whereas micro-level research analyzes the firm ownership, leadership style, and internal resources and their effects on firm green innovation [10,11,12].

In the current digital economy, digital transformation has become the focus for both policymakers and enterprises. Digital transformation is characterized as the integration of digital transformative information technology into organizational practices, resources and capabilities, business processes, and market strategies [15,16]. This new form of technology has not only drastically changed modes of production, business models, organizational forms, and consumer behaviors and expectations but also contributed substantially to reorganizing global resources, as well as reshaping the global economic structure [17]. Prior research on digital transformation mostly focused on corporate strategies, sustainable development, and economic effectiveness [18,19]. For example, scholars have found that digital transformation can increase an organization’s overall factor productivity [20,21]. However, the effect of digital transformation on green innovation in MNEs remains unclear and underexplored. The few studies exploring the effect of digital transformation on green innovation can be classified into two categories: theoretical analysis and case studies, and quantitative research. Although these studies provide some insights and inspirations for understanding digitally driven innovation behaviors, inadequacies remain.

First, despite some research indicating that, in the global digital economy, digital transformation is the key path to sustainable competitiveness for MNEs, most studies in this area focus on theoretical analysis and case discussion [22,23,24], with only a few studies empirically examining the relationship between digital transformation and green innovation performance [25]. Second, in measuring digital transformation, previous research primarily relied on the enterprise’s digital information sharing platform [26] or applied digital technologies [27], which makes it challenging to fully capture the extent of digital transformation and may result in some biases. In addition, prior studies overlooked the digital transformation of MNEs in emerging markets, especially in emerging nations, where firms face significant “outsider disadvantages” or “liability of foreignness” in the course of achieving internationalization. This disadvantage hinders the integration of knowledge and information in host markets, which imposes obstacles to their technological catch-up. To fill the gaps in existing research, this study seeks to address the following questions:

RQ1. What is the relationship between digital transformation and the green innovation performance of MNEs?

RQ2. What are the theoretical mechanisms behind the relationship between digital transformation and the green innovation performance of MNEs?

RQ3. Is the relationship between digital transformation and green innovation performance moderated by the MNEs’ internationalization degree and ownership characteristics?

This study makes three contributions. First, this study advances the theoretical connotations of digital transformation as well as the green innovation performance of MNEs from the perspective of organizational learning theory. In particular, this study offers insight into the mediating role of MNEs’ absorptive capacity, which helps explain the underlying mechanism of the effect of digital transformation on the green innovation performance of MNEs.

Second, this study demonstrates the significant moderating impacts of the extent of internationalization of MNEs and the characteristics of ownership. That is, MNEs operating overseas can formulate specific internationalization strategies to achieve greater innovative progress. Further, the findings not only provide deeper insights in the realm of digitalization but also expand the literature on the effects of MNEs’ green innovation performance in different settings.

Third, the majority of emerging economies are in the midst of a crucial phase of development and expansion. Digitalization can greatly contribute to the restructuring of developing industries and improve the overall productivity of emerging economies. The advancement of digital transformation can be characterized as a technology-enabled and market-based strategic approach; once implemented, the system can be efficient and self-sustainable. MNEs operating abroad rely on the joint influence of digitalization and innovation, which are decisive for a firm’s future growth and sustainability in the global market. Therefore, this study aims to fill the gap in research on digital transformation and green innovation performance. The research results can offer significant insights and guidance for multinational firms aiming to foster and attain a greater green innovation ecosystem.

The remainder of this study is structured as follows. Section 2 briefly illustrates organizational learning theory and develops the research hypotheses. Section 3 presents the research design, including information on the data sample, the measurement of variables, descriptive statistics, and a correlation analysis of all variables. The empirical results are presented in Section 4, and the findings are discussed in Section 5. Finally, Section 6 provides this study’s contributions and implications, along with its limitations and directions for future research.

2. Theoretical Framework and Hypotheses

2.1. Organizational Learning Theory

Organizational learning theory originated from Argyris and Schon’s [28] exploration and elaboration of this notion, which is defined as the process of categorizing, processing, and analyzing different information within an organization from a management perspective. Later, scholars such as Huber [29] and Argyris [30] analyzed organizational learning in depth based on different research contexts and perspectives. Organizational learning theory has become an imperative research topic in many fields, such as sociology, economics, and management. Early organizational learning theories emphasized that organizations could use their prior experiences to identify and acquire relevant knowledge, and that this knowledge is subsequently stored in the form of tacit knowledge in the organizational memory, which can later be processed by the organization to guide its future strategic decisions and actions. Organizational learning is often regarded as a dynamic capability that can determine a firm’s overall performance and sustainable competitiveness on the global market [31].

However, in the era of economic and digital globalization, the market setting has changed dramatically, and the external environment is dynamic and volatile. The survival and development of firms that solely rely on “in-house” learning have become increasingly challenging [32]. Therefore, firms are placing more emphasis on crossing organizational borders and connecting with external entities; this is especially the case for MNEs in emerging economies that want to “take the lead” through innovation. Previous studies on organizational learning theory mainly focused on firms’ engagement and interaction with external subjects, which constitute key intermediaries for obtaining external knowledge and ideas [33,34]. Therefore, developing the ability to identity, acquire, and modify various types of knowledge in the global marketplace becomes an essential factor for the success of MNEs in emerging markets. Therefore, this study explores the effects of digital transformation on the innovation of MNEs in emerging economies from the viewpoint of organizational learning.

2.2. Digital Transformation and Green Innovation Performance of MNEs

Digital transformation refers to a series of processes in which MNEs use transformative information technology to implement major changes in organizational practices, resources and capabilities, business processes, and market strategies [16]. With the arrival of the digital economy, the phenomenon of digitalization has emerged. In addition to promoting information sharing and optimization, this process can strengthen the overall level and quality of green innovation.

This study suggests that the digital transformation of MNEs can improve their innovation performance through several mechanisms. First, digital transformation assists MNEs in acquiring and using various green-related information from host markets, as well as facilitates the analysis and mining of various information sources based on information technology [35]. This process allows MNEs to identify deficiencies in their own technology—by comparing their knowledge bases with those in overseas markets, they can update their technology and further innovate. Moreover, digital transformation can also help MNEs overcome the “outsiders’ disadvantage” in different ways, such as reducing the negative effects of the institutional and cultural distances between the home and host markets. As a result, MNEs can explore overseas markets to obtain relevant information and resources and improve their ability to process various types of information.

Second, digital transformation strengthens the link between MNEs and their stakeholders in overseas markets [36]. For example, MNEs can maintain close contact with their stakeholders in overseas markets via the Internet of Things, blockchain, and crowdsourcing and crowdfunding systems and continuously enhance trust in these relationships [37], thereby ensuring continual access to overseas market information. The trust in these relationships is a precondition for close cooperation between multinational enterprises and their stakeholders. This diminishes the risk and cost of R&D cooperation in host markets, thereby promoting innovation in MNEs.

Third, digital technologies facilitate information and knowledge dissemination and exchange in global innovation networks [38], which reduces the level of information asymmetry in global networks. In addition, digital facilities and technologies can also reinforce communication and engagement within MNEs, improving the efficiency of collaboration between subsidiaries and parent companies [39]. Digital technology can enhance collaboration and provide mutual benefits for both domestic and foreign innovation networks of MNEs. Based on the above theoretical analysis, this study puts forward the following hypothesis:

H1. 

Digital transformation of MNEs can positively affect their green innovation performance.

2.3. The Mediating Role of Absorptive Capacity

An enterprise’s absorptive capacity refers to its ability to recognize, obtain, incorporate, convert, and utilize knowledge of the external environment and transform it directly into economic output [40]. Through absorption, MNEs can combine existing and new knowledge, thus contributing to innovation [41,42]. This study hypothesizes that absorptive capacity serves as a mediator between digital transformation and green innovation performance for the reasons described below.

First, the digital transformation of MNEs provides them with valuable information from domestic and foreign markets, enabling them to fully comprehend the needs of stakeholders at home and abroad. In other words, digital technology can be utilized to transform massive routine data into accurate and effective information, which will ultimately improve the efficiency of MNEs in processing big data information [43]. Digital transformation also allows MNEs to accurately obtain necessary resources and capabilities according to the current circumstances of the dynamic competitive environment in domestic and foreign markets. This transformation allows MNEs to better align their resources and capabilities, which is conducive to the absorption and conversion of relevant information and knowledge.

Second, the digital transformation of MNEs can enhance their overall ability to identify a broad range of information and resources, making it easier to acquire, select, absorb, and transform key information and tacit knowledge. Digital transformation disrupts the traditional management model of MNEs. Scholars have revealed that digital technologies facilitate knowledge and resource transfer between entities and enhance their overall information-processing skills. This enabled companies to strengthen their internal regulatory structure, production approach, and operational mechanism and reduce the cost of internal management [44]. In addition, digital transformation can break the restraints of traditional factor boundaries, enabling sufficient information sharing between MNEs and their stakeholders, reducing information asymmetry between different stakeholders, and facilitating the absorption and transformation of various information and knowledge [45].

Third, many studies have discussed the effect of absorptive capacity on the green innovation performance of MNEs. Researchers posit that absorptive capacity can help MNEs distinguish and obtain valuable information in the global marketplace and encourage knowledge exchange and information sharing between parents and subsidiaries of MNEs, positively affecting their green innovation performance [45]. Based on the above theoretical analysis, this study proposes the following:

H2. 

Absorptive capacity can positively mediate the relationship between digital transformation and green innovation performance of MNEs.

2.4. The Moderating Role of the Degree of Internationalization

The degree of internationalization of a multinational enterprise (MNE) reflects, to some extent, the overseas resources it possesses and the extent of competition it confronts. This study hypothesizes that the degree of MNE internationalization positively moderates the relationship between digital transformation and absorptive capacity for the reasons outlined below.

First, when increasing their extent of internationalization, MNEs are able to master and acquire more resources from host markets [46], creating a resource foundation to support innovation activities. In other words, a high degree of internationalization indicates that MNEs have more resources in foreign markets, which can help them identify and evaluate existing technologies through digital technology. As a result, these MNEs can increase R&D investment more strategically, achieve breakthroughs, and thus enhance their absorptive capacity.

Second, a high degree of internationalization means that MNEs face intense competition in overseas markets [47]. To gain a competitive lead in the fierce international marketplace, it is essential to innovate and surpass the limits of existing technology. This provides MNEs with the chance to evaluate the innovation ecosystem, make predictions, and avoid the risk of being isolated from pertinent information [48]. Hence, a greater extent of MNE internationalization facilitates the use of digital technology to identify shortcomings in current technologies and create new technologies by means of R&D investments.

Third, engaging in operational activities in various foreign markets can help counterbalance the institutional shortcomings of the home market. That is, a high degree of internationalization provides MNEs with greater institutional convenience and benefits from a more wide-ranging overseas market. Thus, a wide geographic reach allows MNEs to find compatible institutional settings in which they can effectively leverage their advantages, improving their acquisition, integration, and absorption of foreign information and knowledge [49]. Thus, this study proposes the following:

H3. 

The degree of internationalization can enhance the positive effect of digital transformation of MNEs on absorptive capacity.

2.5. The Moderating Role of the State Ownership

This study hypothesizes that the impact of digitization on absorptive capacity is also affected by the characteristics of ownership; i.e., state ownership moderates the relationship between digital transformation and absorptive capacity. The reasons for this are as follows.

First, compared with private companies, state-owned MNEs have a closer relationship with the government, which makes them more likely to have access to the government’s resources and support [50]. It has been argued that the success of many state-owned MNEs in overseas markets can be attributed to certain government-related resources [51]. Due to lenient funding and budgets, state-owned MNEs often attach more importance to strategic financial models than to short-term profit-maximizing strategies [52]. This provides MNEs with the guarantee of continuous innovation activities, strengthening their motivation to engage in long-term investments and increasing their willingness to understand their own technological shortcomings and deficiencies.

Second, compared with private enterprises, state-owned MNEs face greater governmental demands and expectations in terms of innovation, as such MNEs contribute considerably to the country’s economic growth [53]. Thus, governments, particularly in developing economies, place high expectations and pressure on state-owned MNEs, given their economic importance, to strengthen their technological capabilities and promote innovation. As a result, state-owned MNEs are pressured to play a proactive role in initiating and executing innovation.

H4. 

The effect of digital transformation on absorptive capacity is stronger in state-owned MNEs compared to private MNEs.

3. Methods

3.1. Sample and Data Collection

The Chinese economy has developed substantially since the country initiated reform and opening-up policies. These policies were influenced heavily by the national strategy of “bringing in and going out,” which has allowed Chinese manufacturing companies to significantly strengthen their innovation ecosystem and make notable progress in innovation. This study focuses on Chinese listed multinational enterprises, which are defined as enterprises with at least one overseas subsidiary, according to [54], as the research objects. To ensure the validity of the research samples, the samples were screened using following criteria: (1) enterprises that had been listed for less than 2 years were excluded from the sample to avoid irregular cash-holding levels following the initial public offering; (2) financial and insurance industries with assets, financial structure, and operational approaches significantly different from those of other industries were excluded; (3) *ST enterprises with poor financial conditions or other abnormal circumstances were excluded; (4) sample enterprises with missing data were excluded. The data on the digital transformation of MNEs came from annual reports disclosed by the enterprises, and the rest of the data came from the CSMAR database. Ultimately, 2423 observations from 2010 to 2019 were included in the analysis.

3.2. Measures

3.2.1. Dependent Variable: Green Innovation Performance

Two approaches are used to measure green innovation performance. One is based on the MNEs’ green patent information, which includes the number of granted or applied patents [55], the number of patent citations [56], etc. The other way to measure innovation performance is to use the sales of new products by the MNEs [57,58]. Since the Chinese government does not require listed enterprises to disclose the sales revenue of new products in their annual reports, many listed enterprises have large amounts of missing data in their annual reports, so this study employed the green patent information published in the annual reports of listed enterprises to measure innovation performance. Green invention patents contain more technical content than other sources, which can better represent the green innovation output of MNEs [59]. In accordance with [59], this study uses the number of granted green invention patents to measure the green innovation performance of MNEs. To eliminate the influence of heteroscedasticity, the number of green invention patents granted to listed MNEs is logarithmic.

3.2.2. Independent Variable: Digital Transformation

Prior studies have attempted to quantify digital transformation using metrics such as the ratio of digital intangible assets. However, digital transformation involves a vast amount of unstructured data, making it difficult to measure by traditional structured techniques accurately. The advantage of text mining method is that the results are intuitive and easy to understand, and they objectively reflect the degree of digital transformation of enterprises, which avoids the interference of the researcher’s subjective judgment to the greatest extent. Annual reports are usually released regularly. Text mining method can capture the slight changes in a firm’s strategic priorities in near real time and dynamically track the progress of the firm’s digital transformation. Additionally, by analyzing the frequency, weight and context of the words related to digitalization in firms’ annual reports, it is possible to precisely measure the extent to which the firm management allocates their attention and strategic resources to digital transformation, which provides a forward-looking on the strategic intent of firms. Thus, text mining based on the annual reports of firms has been increasingly used as an effective method for quantifying firms’ digital transformation efforts. Following the study of [21], the text-mining method was used in this study to assess the digital transformation of MNEs. To ensure the reliability and relevance of the text mining approach, this study follows a rigorous three-step process. The construction of the digital transformation indicator follows a three-step process. First, relevant keywords are identified and screened based on their prevalence in academic literature and industry practices (see

Table 1. The thesaurus of key words of digital transformation.

Classification	Key Words
Artificial intelligence technology	artificial intelligence, machine learning, neural networks, biometrics, face recognition, deep learning, natural language processing, image recognition, automatic speech recognition, sentiment analysis, human–computer interaction, intelligent manufacturing, flexible manufacturing, automation, 3D printing, robotics, active manufacturing, intelligent manufacturing, intelligent enterprise, intelligent terminal, intelligent identification
Digital technology	digital technology, digitization, digital twin, digital economy, big data, data mining, data empowerment, data assets, data visualization, cloud computing, cloud platform, cloud manufacturing, internet of things, blockchain
Internet information technology	internet, internet plus, industrial internet, informatization, information technology, information, and communication technology

). Second, these keywords are extracted from the MD&A sections of listed firms’ annual reports on a yearly basis, allowing for the tracking of evolving terminology and trends. The annual frequency of these keywords is quantified to measure the emphasis a firm place on digital transformation. Finally, the raw frequency count is transformed by applying the natural logarithm to standardize the measure and mitigate potential skewness in its distribution.

3.2.3. Mediating Variable: Absorptive Capacity

R&D is a decisive element in improving the absorption, filtering, and storage of external knowledge of an enterprise, and an enterprise’s absorption capacity rises with increased R&D efforts [60]. However, due to the extensive variation in the investment and degree of R&D amongst various companies, in line with [61], this study measures the absorptive capacity of MNEs using R&D intensity, measured as the natural logarithm of the ratio of R&D investment to the total assets.

3.2.4. Moderating Variables: State Ownership and Internationalization Degree

The degree of MNE internationalization is measured by the natural logarithm of the number of foreign countries in which enterprises have investments, following the study of [56].

State ownership is measured by a dummy variable—if the MNE is state-owned, then it is coded as 1; otherwise, it is 0.

3.2.5. Control Variables

To control the influence of other potential variables, the following variables are controlled in our study. Firm age is measured by the natural logarithm of the number of years for which the MNE has been established. Firm size is measured by the total number of employees in the MNE, net profit is measured by the total net profit of the MNE, and overseas income is measured by the total overseas sales revenue of the MNE, all natural-logarithm-transformed. CEO openness to change is measured by three indicators: CEO tenure, age, and education level; this approach follows the study of [62]. Ownership concentration is measured by the proportion of the largest shareholder of the MNE. The shareholding ratio of managers is measured by the proportion of managers’ shareholding ratios in MNEs. The proportion of independent directors is measured by the proportion of the number of independent directors to the total number of board members. The dual role of the CEO is measured as “0–1”—if the CEO is also the chair of the MNE, then it is coded as 1; otherwise, it is coded as 0. Global dynamic capability is measured by three dimensions: international human capital, international social capital, and international managerial cognitions. This approach follows the studies of [63,64]. The asset–liability ratio is the proportion of total assets to total liabilities of an enterprise, which is natural-logarithm-transformed. Manufacture is measured as “0–1”; if the MNE is a manufacturer, then it is coded as 1, and 0 otherwise.

3.3. Descriptive Statistics and Correlation Analysis

Table 2. Descriptive statistics and correlation analysis.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)	(13)	(14)	(15)	(16)	(17)
(1) Innovation performance	1
(2) Digital transformation	0.065 *	1
(3) Absorptive capacity	0.072 **	0.262 ***	1
(4) State-owned MNEs	−0.011	−0.093 ***	−0.007	1
(5) Internationalization Degree	0.047 *	0.043 *	0.146 ***	0.000	1
(6) MNE age	0.007	−0.110 ***	−0.099 **	0.334 ***	−0.024	1
(7) MNE size	−0.049 *	−0.008	0.052	0.311 ***	0.279 ***	0.214 ***	1
(8) Net profit	−0.011	0.057 **	0.030	0.187 ***	0.270 ***	0.103 ***	0.637 ***	1
(9) Overseas income	−0.023	−0.096 ***	0.024	0.248 ***	0.386 ***	0.132 ***	0.592 ***	0.413 ***	1
(10) CEO openness to change	−0.007	0.014	0.008	0.030	−0.053 **	−0.068 ***	−0.075 ***	−0.051 *	−0.067 ***	1
(11) Ownership concentration	0.014	−0.115 ***	0.017	0.165 ***	0.008	0.034	0.099 ***	0.091 ***	0.062 **	0.027	1
(12) The shareholding ratio of management	0.034	0.127 ***	0.107 ***	−0.404 ***	−0.020	−0.231 ***	−0.251 ***	−0.176 ***	−0.199 ***	−0.125 ***	−0.057 **	1
(13) Proportion of independent directors	−0.035	0.030	0.023	−0.006	0.014	0.026	−0.055 **	−0.025	−0.015	−0.008	0.082 ***	0.081 ***	1
(14) Dual role of CEO	−0.004	0.114 ***	0.087 **	−0.286 ***	0.011	−0.126 ***	−0.094 ***	−0.092 ***	−0.045 *	−0.197 ***	0.004	0.035	0.092 ***	1
(15) Global dynamic capability	−0.038	0.007	0.053	−0.047 *	0.202 ***	−0.026	0.042 *	0.020	0.149 ***	0.029	−0.127 ***	0.028	−0.013	0.017	1
(16) The ratio of asset liability	0.018	0.098 ***	0.109 ***	−0.260 ***	−0.141 ***	−0.194 ***	−0.448 ***	−0.219 ***	−0.369 ***	0.063 **	−0.072 ***	0.255 ***	0.075 ***	0.122 ***	0.026	1
(17) Manufacture	0.046 *	−0.250 ***	0.045	−0.012	−0.015	−0.036	0.068 ***	−0.070 ***	0.181 ***	−0.075 ***	0.098 ***	0.018	0.001	0.072 ***	−0.004	0.007	1
Mean	0.230	1.934	0.031	0.222	0.971	1.405	8.129	19.112	19.762	−0.204	3.408	1.194	38.119	0.383	0.369	1.342	0.786
Standard deviation	0.775	1.416	0.027	0.416	0.608	0.060	1.150	1.451	2.029	0.689	0.432	1.405	5.709	0.486	0.482	0.428	0.410

Note: * p < 0.05, ** p < 0.01, *** p < 0.001.

illustrates all variables’ descriptive statistics and the correlations between them. The correlation between the digital transformation and green innovation performance of MNEs is positive and significant (γ = 0.065, p < 0.05), fulfilling a prerequisite for exploring the causal relationship between digital transformation and green innovation performance; the correlation between digital transformation and absorptive capacity is positive and significant (γ = 0.262, p < 0.001). The correlation coefficient between absorptive capacity and innovation performance is 0.072 and significant (p < 0.01), which indicates that absorptive capacity may positively mediate the relationship between digital transformation and innovation performance, providing a premise for the subsequent analysis of the mediating effect. In addition, the VIF values of the variables in this paper are all less than 10, indicating that there is no serious collinearity among the variables. Therefore, a subsequent data analysis can be conducted to test the causal relationship between the digital transformation and green innovation performance of MNEs and further assess the mediating role of absorptive capacity and the moderating role of the degree of internationalization and state ownership.

4. Results

4.1. Main Effects and Mediation Effects Tests

To further investigate and analyze the causal link between the digital transformation and green innovation performance of MNEs and reveal the mediating effect of absorptive capacity, Sata17 was used to conduct regression analysis, and the results are shown in

Table 3. The results of regression analysis.

Variables	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
	Innovation Performance	Innovation Performance	Absorptive Capacity	Innovation Performance	Absorptive Capacity	Absorptive Capacity
Firm age	0.125	0.195	−0.032 *	0.805 †	−0.034 *	−0.027 †
	(0.440)	(0.684)	(−2.266)	(1.669)	(−2.370)	(−1.933)
Firm size	−0.052 *	−0.056 *	0.003 **	−0.102 *	0.003 **	0.003 **
	(−2.269)	(−2.441)	(2.825)	(−2.448)	(2.719)	(2.854)
Net profit	0.022	0.016	−0.000	0.042	−0.000	−0.001
	(1.413)	(1.062)	(−0.142)	(1.602)	(−0.208)	(−0.756)
Overseas income	0.001	0.005	−0.000	0.014	−0.000	−0.001
	(0.128)	(0.482)	(−0.235)	(0.688)	(−0.320)	(−1.340)
CEO openness to change	−0.009	−0.009	0.000	−0.044	0.000	0.000
	(−0.355)	(−0.376)	(0.250)	(−1.060)	(0.150)	(0.356)
Ownership concentration	0.057	0.062	0.001	0.075	0.002	0.001
	(1.456)	(1.598)	(0.676)	(1.111)	(0.777)	(0.758)
The shareholding ratio of manager	0.022	0.020	0.001	0.039	0.001	0.001
	(1.531)	(1.420)	(0.955)	(1.514)	(1.307)	(0.831)
Proportion of independent directors	−0.005	−0.005 †	−0.000	−0.002	−0.000	−0.000
	(−1.634)	(−1.714)	(−0.837)	(−0.336)	(−0.840)	(−0.956)
Dual role of CEO	−0.059	−0.068 †	0.001	−0.104	0.001	0.001
	(−1.461)	(−1.681)	(0.294)	(−1.504)	(0.398)	(0.413)
Global Dynamic Capability	−0.055	−0.056	0.001	−0.076	0.001	−0.000
	(−1.586)	(−1.622)	(0.348)	(−1.411)	(0.557)	(−0.247)
The ratio of asset liability	0.012	0.012	0.006 *	−0.047	0.006 *	0.006 *
	(0.256)	(0.271)	(2.443)	(−0.562)	(2.321)	(2.430)
Manufacture	−0.606 †	−0.585 †	−0.021	−0.551	−0.018	−0.021
	(−1.729)	(−1.674)	(−1.493)	(−1.168)	(−1.323)	(−1.502)
Industry	control	control	control	control	control	control
Year	control	control	control	control	control	control
Constant	0.286	0.169	0.018	−1.100	0.023	0.036
	(0.486)	(0.287)	(0.542)	(−0.971)	(0.687)	(1.081)
Digital transformation		0.052 ***	0.004 ***	0.054 *	0.003 ***	0.002 *
		(3.756)	(5.528)	(2.250)	(3.715)	(2.536)
Absorptive capacity				2.290 *
				(2.146)
State ownership × Digital transformation					0.004 **
					(2.869)
State ownership					−0.004
					(−1.293)
Internationalization × Digital transformation						0.002 *
						(2.063)
Internationalization						0.002
						(0.732)
No. of observations	2423	2423	1089	1089	1089	1089
R2	0.067	0.072	0.197	0.096	0.204	0.210
adj. R2	0.035	0.040	0.147	0.039	0.153	0.159
F value	2.09	2.25	3.92	1.68	3.98	4.12

Note: t statistics in parentheses; ^† p < 0.1; * p < 0.05, ** p < 0.01, *** p < 0.001.

. Model 1 is the benchmark model and comprises only the control variables. The results of model 2 indicate that digital transformation can positively affect the green innovation performance of MNEs (β = 0.065, p < 0.001), which supports H1.

The regression results regarding the mediating role of absorptive capacity are as follows: First, according to model 2, the digital transformation of MNEs can positively affect their green innovation performance. Then, as shown in model 3, digital transformation positively affects absorptive capacity (β = 0.004, p < 0.001), which in turn can improve the green innovation performance of MNEs (β = 2.290, p < 0.05), as shown in model 4. Thus, absorptive capacity can positively mediate the effect of digital transformation on the green innovation performance of MNEs. H2 is supported. In addition, according to model 4, the relationship between digital transformation and green innovation performance is positive and significant (β = 0.054, p < 0.05); it can be further inferred that absorptive capacity plays a partial mediating role.

The above testing process is mainly based on the stepwise regression method, which is the least powerful method for testing mediating effects. It is difficult to detect a significant mediating effect by using the stepwise regression coefficient if the effect is weak. On the other hand, if the existence of the intermediary effect is tested, then the low strength of the stepwise method is no longer a problem. To further test the significance of the intermediary effect, the non-parametric Bootstrap method for repeated sampling (referring to the study by [65] was applied to generate multiple “new samples”, and statistical analysis of the “new samples” was performed to obtain a more accurate parameter estimation. The results are shown in

Table 4. Mediating effect test (Bootstrap method).

	Coefficient	Standard Error	Z Value	p Value	95% Confidence Interval
Indirect effect	0.009	0.002	4.32	0.000	[0.005, 0.013]
Direct effect	0.054	0.013	4.14	0.000	[0.028, 0.079]

(the number of bootstrap samples is set to 5000). It can be seen that the 95% confidence interval for indirect effects (0.005, 0.013) does not contain 0, indicating that the mediating effect of the absorptive capacity is significant and positive. Moreover, the 95% confidence interval of the direct effect (0.028, 0.079) does not contain 0, indicating that the direct effect is significant, which again indicates that the absorptive capacity plays a partial mediating role in the relationship between the digital transformation and green innovation performance of MNEs.

4.2. Moderating Effects Tests

In addition, according to model 6 in Table 3, the interaction between the degree of internationalization and the digital transformation of MNEs is positive and significant (β = 0.002, p < 0.05), which supports H3. A high degree of MNE internationalization can enhance the positive effect of digital transformation on absorptive capacity. According to model 5, the interaction between state ownership and digital transformation is positive and significant (β = 0.004; p < 0.01), which supports H4. Thus, the positive effect of digital transformation on absorptive capacity may be stronger in state-owned MNEs than in private MNEs.

4.3. Robustness Tests

To further test the robustness of the research findings, the following tests and analyses were also carried out.

The explanatory variables are lagged. To avoid the potential endogeneity problem caused by mutual causality, the explanatory variables in the t period and the explained variables in the t + 1 period were used for regression analysis, and the results are shown in

Table 5. The results of robustness test (lagged one year).

Variables	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
	Innovation Performance	Innovation Performance	Absorptive Capacity	Innovation performance	Absorptive Capacity	Absorptive Capacity
Absorptive capacity				2.054 † (1.228)
Digital transformation		0.049 *	0.003 ***	0.052 †	0.001 †	0.002 †
		(0.020)	(0.001)	(0.031)	(0.001)	(0.001)
State ownership × Digital transformation					0.004 **
					(0.001)
State ownership					−0.003
					(0.003)
Internationalization × Digital transformation						0.001
						(0.001)
Internationalization						0.005 *
						(0.002)
Firm age	0.016	0.072	−0.042 **	0.655	−0.045 **	−0.033 *
	(0.398)	(0.398)	(0.015)	(0.612)	(0.015)	(0.015)
Firm size	−0.092 **	−0.094 **	0.002	−0.170 **	0.002	0.002
	(0.033)	(0.033)	(0.001)	(0.053)	(0.001)	(0.001)
Net profit	0.052 *	0.045 *	0.000	0.081 *	0.000	−0.001
	(0.022)	(0.022)	(0.001)	(0.033)	(0.001)	(0.001)
Overseas income	0.019	0.025	−0.000	0.030	−0.000	−0.001
	(0.016)	(0.016)	(0.001)	(0.027)	(0.001)	(0.001)
CEO openness to change	0.005	0.007	0.000	0.015	0.000	0.001
	(0.035)	(0.035)	(0.001)	(0.053)	(0.001)	(0.001)
Ownership concentration	0.064	0.069	0.001	0.077	0.001	0.001
	(0.056)	(0.056)	(0.002)	(0.088)	(0.002)	(0.002)
The shareholding ratio of manager	0.044 *	0.042 *	0.000	0.063 †	0.001	0.000
	(0.020)	(0.020)	(0.001)	(0.032)	(0.001)	(0.001)
Proportion of independent directors	−0.005	−0.005	−0.000	−0.002	−0.000	−0.000
	(0.004)	(0.004)	(0.000)	(0.007)	(0.000)	(0.000)
Dual role of CEO	−0.110 †	−0.119 *	0.002	−0.181 *	0.002	0.002
	(0.057)	(0.057)	(0.002)	(0.088)	(0.002)	(0.002)
Global Dynamic Capability	−0.116 *	−0.119 *	−0.001	−0.149 *	−0.001	−0.002
	(0.047)	(0.047)	(0.002)	(0.066)	(0.002)	(0002)
The ratio of asset liability	0.026	0.035	0.007 **	−0.136	0.007 **	0.007 **
	(0.063)	(0.063)	(0.003)	(0.106)	(0.003)	(0.002)
Manufacture	0.096	0.087	−0.007	−0.131	−0.006	−0.003
	(0.263)	(0.262)	(0.013)	(0.427)	(0.013)	(0.013)
Industry	control	control	control	control	control	control
Year	control	control	control	control	control	control
Constant	−0.004	−0.135	0.047	−0.313	0.056	0.058†
	(0.851)	(0.851)	(0.035)	(1.531)	(0.034)	0.034
No. of observations	1423	1423	625	766	625	625
R2	0.091	0.095	0.269	0.130	0.286	0.293
adj. R2	0.042	0.046	0.197	0.056	0.213	0.220
F value	1.87	1.93	3.73	1.75	3.92	4.04

Note: Standard error in parentheses; ^† p < 0.1; * p < 0.05, ** p < 0.01, *** p < 0.001.

. First, model 1 is the benchmark model, and model 2 shows the relationship between the digital transformation and green innovation performance of MNEs. The results show that digital transformation has a positive impact on green innovation performance (β = 0.049; p < 0.05), supporting hypothesis H1. Second, from model 3, digital transformation positively affects absorptive capacity (β = 0.003; p < 0.001). Finally, the results of model 4 indicate that absorptive capacity positively affects green innovation performance (β = 2.054; p < 0.1), and the relationship between digital transformation and green innovation performance is positive and significant (β = 0.052; p < 0.1). The results of stepwise regression reveal that absorptive capacity positively mediates the relationship between digital transformation and green innovation performance and plays a partial mediating role, further supporting hypothesis H2. Models 5 and 6, respectively, show the moderating effects of state ownership and the degree of internationalization. According to model 5, the interaction term between state-owned MNEs and digital transformation is positive and significant (β = 0.004; p < 0.01), suggesting that H4 is supported. According to model 6, the interaction term between the degree of internationalization and digital transformation is positive but not significant (β = 0.001; p > 0.1), indicating that H3 is not supported. Therefore, after lagging all explanatory variables by one period, nearly all hypotheses except H3 are supported, which indicates that the results have robustness.

Heckman’s Two-stage Method: The prior analysis demonstrated that digital transformation can promote green innovation in MNEs to a certain extent. However, MNEs that want to improve their green innovation performance are more likely to engage in active digital transformation for its potential benefits, which may have introduced a bias due to the self-selection of research samples. To test the influence of sample self-selection bias, Heckman’s two-stage method was adopted.

Heckman’s first-stage regression involves setting a dummy variable to determine whether the level of digital transformation of MNEs is higher than the mean value. Then, a probability equation is constructed as an explanatory variable to predict the probability of MNEs disclosing information pertaining to digital transformation, and the inverse Mills ratio is calculated. On the basis of controlling the original control variables, the number of domain names owned by the region is controlled as an exogenous variable, which is an important aspect of the regional digital infrastructure and is closely related to the digital transformation of enterprises. This variable, which mainly reflects the status of digital infrastructure in a region, has a weak impact on an enterprise’s green innovation activities and meets the externality criteria. In the second regression stage, the inverse Mills ratio (IMR) obtained in the first stage is incorporated into the model. The regression results are shown in

Table 6. The results of Heckman Analysis.

Variables	Innovation Performance
	Model 1
IMR	−0.272
	(0.290)
Digital Transformation	0.046
	(0.030)
Firm age	0.306
	(0.436)
Firm Size	−0.104 **
	(0.037)
Net Profit	0.040
	(0.027)
Overseas Incomes	0.018
	(0.018)
CEO openness to change	−0.055
	(0.040)
Ownership concentration	0.063
	(0.062)
The shareholding ratio of manager	0.033
	(0.022)
Proportion of independent directors	−0.008 †
	(0.005)
Dual role of CEO	−0.135 †
	(0.075)
Global Dynamic Capability	−0.078
	(0.054)
The ratio of asset liability	0.043
	(0.069)
Manufacture	−0.680
	(0.469)
Industry	control
Year	control
Constant	0.103
	(0.951)
N	1252
R2	0.078
adj. R2	0.026
F-value	1.50
Root MSE	0.833

Note: Standard error in parentheses; ^† p < 0.1; ** p < 0.01.

. Model 1 shows that the coefficient of the IMR is non-significant (β = −0.272; p > 0.05), indicating that no serious sample self-selection bias exists in this study.

Instrumental variable analysis. This study used two-stage least squares (2SLS) instrumental variable regression to address potential endogeneity concerns. The mean value of digital transformation for all firms in the same industry and province in a year is chosen as the instrument. The 2SLS regression results are presented in

Table 7. The results of 2 SLS Analysis.

Variables	First Stage	Second Stage
	Digital Transformation	Green Innovation Performance
Instrumental variable	0.952 ***
	(0.040)
Digital Transformation		0.002 *
		(0.001)
Controls	Yes	Yes
Industry	Yes	Yes
Year	Yes	Yes
Constant	0.393	0.026
	(1.240)	(0.033)
N	1089	1089
R2	0.593	0.194
F-value/Wald χ2	23.33	239.65
Root MSE	0.910	0.024

Note: Standard error in parentheses; * p < 0.05, *** p < 0.001.

. In the first stage, the instrumental variable is shown to be positively correlated with the digital transformation of MNEs (β = 0.952, p < 0.001). In the second stage, the results indicate that the coefficient for digital transformation remains positive and statistically significant (β = 0.002, p < 0.05). The results of weak instrumental variable tests show that the Cragg–Donald Wald F statistic is 522.871, significantly exceeding the Stock–Yogo 10% maximal IV size critical value of 16.38, which means that the instruments are strong.

5. Discussion

From the perspective of organizational learning, this study deepens our understanding of the drivers of MNEs’ green innovation performance in the digital era. First, we explored the effect of digital transformation on the green innovation performance of MNEs and identified the boundary conditions of the degree of internationalization and state ownership. The findings suggest that digital transformation positively influenced MNEs’ green innovation performance. For MNEs from emerging economies, which are often at a technological disadvantage in international markets, digital transformation provides a path for overtaking competition through technological means. Digital transformation contributes to MNEs’ identification of and access to information in the global marketplace, enabling information sharing between MNEs and host-country stakeholders [36]. Further, digital transformation also allows MNEs to extract new knowledge from massive quantities of data, therefore providing information and knowledge support for MNEs’ green innovation activities.

Furthermore, the results also revealed the theoretical mechanisms underlying the relationship between the digital transformation and green innovation performance of MNEs; i.e., digital transformation improves the green innovation performance of MNEs by enhancing their absorptive capacity. This signifies that digital transformation provides a path through which MNEs can enhance the integration and absorption of information and knowledge at home and abroad [43], which in turn improves their ability to reprocess information and knowledge. Moreover, the transformation and absorption of information from an external environment are essential for the success of green innovation activities [41,42]. This encourages MNEs to create new green technologies and business models with the aim of strengthening their competitive advantages in the global market.

Finally, the findings suggest that the extent of MNE internationalization and state ownership can moderate the relationship between the digital transformation of MNEs and their absorptive capacity. Specifically, State ownership positively moderates the relationship between digital transformation and absorptive capacity; that is, the positive influence of digital transformation on absorptive capacity is stronger in state-owned MNEs than in private MNEs. The degree of internationalization can positively moderate the relationship between digital transformation and absorptive capacity [66,67]. This proves that the impact of digital transformation on absorptive capacity is stronger for MNEs with a high degree of internationalization. But interestingly, this study finds that the moderating effect of internationalization degree may diminish as time goes by, which means that the moderating effect of internationalization degree is more likely to take effect simultaneously rather than lag by one period. This finding reveals that the positive effects of internationalization may not last for a long time, which is different with state ownership.

6. Contributions, Implications, and Limitations

6.1. Theoretical Contributions

Our research has several theoretical implications. First, by investigating the effect of digital transformation on green innovation performance in the digital era, our study expands the literature on facilitators of MNEs’ green innovation performance. Previous studies mainly investigated the contributing factors of green innovation performance from the viewpoint of a company’s motivating and internal factors [7]. However, the influence of digitalization on green innovation performance was not fully explored. Our results expand the body of literature on green innovation performance and contribute to the establishment of a theoretical relationship between these two variables.

Second, our study demonstrates the mediating role of MNEs’ absorptive capacity, through which digital transformation can positively influence green innovation in MNEs. The findings advance our interpretation of the mechanism underpinning the influence of digital transformation on green innovation performance. Whereas some prior research focused solely on the direct relation between digital transformation and green innovation performance and neglected its mediating and indirect effects [57], our research extends beyond the direct links by explicitly examining the underlying mechanism between these two variables. Furthermore, this study offers new directions for future research on the relationships between the digital economy and green innovation performance from an emerging economy stand-point.

Third, this study revealed the boundary effects of MNEs’ degree of internationalization and state ownership. The findings indicate that a high degree of internationalization and state ownership enhanced the relationship between the digital transformation and green innovation performance of MNEs. These two moderating variables can better define the complex link between digital transformation and green innovation performance. This suggests that under certain conditions, MNEs can more effectively assimilate and integrate external knowledge and information through digitalization, leading to the enhanced implementation of green innovation in firms.

6.2. Practical Implications

The results of this study also have the following managerial implications for the operation and growth of MNEs. First, to remain competitive in the current global digital economy, MNEs should actively pursue digitalization and embrace various digital technologies, such as cloud computing, artificial intelligence, and blockchain. It is crucial to acknowledge the potential value of digital technology. MNEs should actively seek to technologically advance and obtain valuable global market information and knowledge by increasing investment in digital technology. In addition, the government should provide MNEs with subsidiaries and technical support, as well as incentives to encourage them to adopt digital transformation and implement green innovation strategies.

Second, MNEs should value their ability to incorporate and absorb new information and knowledge and increase investment in R&D. Although digital transformation provides a way for MNEs to access a variety of information and knowledge in the global market [17], it is more important that they integrate and transform the information and knowledge, as well as create their own knowledge base and system. The strength of their absorptive capacity directly determines the effects of green innovation in MNEs. Therefore, if MNEs in emerging economies want to overcome the technological restrictions of host countries and achieve technological breakthroughs in the current digital economy, they need to enhance their own absorption capacity while prioritizing the advancement of enterprise digital technology.

Third, it is essential to take into consideration the moderating factors of the degree of internationalization and state ownership, as these two factors greatly impact the relationship between MNE’s digital transformation and absorptive capacity. For instance, managers need to consider the scope of internationalization strategy and the characteristics of their ownership. At a policy level, the government can leverage the power of tax breaks in order to motivate MNEs to enlarge their breadth of internationalization. In conclusion managers can weigh the particular circumstances of the firm and establish strategies that is both favorable and suitable for its innovation development.

6.3. Limitations and Future Research

Although this study has obtained numerous valuable findings and enriched the existing literature, limitations still exist. First, Chinese listed MNEs were the primary focus of our study, and we thoroughly contextualized our theory within this framework. We call for future research to assess and enhance our theory in other settings and emerging entities. In addition, the implementation of green innovation activities is often a complex and time-consuming systematic project. Limited by research data, our study measured green innovation using only the number of green patent applications. Future studies can consider utilizing a broader range of measurement techniques, such as those in the social, environmental, and economic domains.

Second, this study used the text-mining method to assess the digital transformation of MNEs. Although this method is widely used, it may overestimate the actual level of digital transformation of firms, because “talking” is always much easier than “doing”. Future studies can design new methods to effectively measure the degree of digital transformation, such as through the application of digital patents. Additionally, absorptive capacity represents a multi-dimensional capability that involves the acquisition, assimilation, transformation, and exploitation of knowledge, and this study may have failed to capture these multi-dimensional characteristics by only using the intensity of R&D to measure absorptive capacity. Thus, future studies can develop more appropriate indicators to comprehensively examine all dimensions of absorptive capacity.

Third, this study examined the effects of digital transformation on green innovation from an organizational learning perspective. However, the relationship between these two variables is very complex and may involve multiple factors. Therefore, subsequent studies can further explore the mechanism of the relationship between them from other theoretical perspectives, such as attention-, human-, or financial resources-based views.