Next Article in Journal
Drone-Assisted Order Picking Problem: Adaptive Genetic Algorithm
Previous Article in Journal
Resilience Through Integration: The Synergistic Role of National and Organizational Culture in Enhancing Market Responsiveness
Previous Article in Special Issue
Deliberate and Emergent Strategic Outcomes for High-Growth IT SME Business Models
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Dual-Dimensional Digital Transformation Systematically Reshapes the U-Curve of Knowledge and Political Distance on Subsidiary Exit

Glorious Sun School of Business & Management, Donghua University, Shanghai 200051, China
*
Author to whom correspondence should be addressed.
Systems 2025, 13(9), 773; https://doi.org/10.3390/systems13090773
Submission received: 31 July 2025 / Revised: 28 August 2025 / Accepted: 2 September 2025 / Published: 3 September 2025
(This article belongs to the Special Issue Business Model Innovation in the Digital Era)

Abstract

In the era of digital business model innovation, multinational corporations face a dual challenge: leveraging digital technologies to overcome institutional barriers while reconfiguring value creation in cross-border operations. Grounded in institutional theory and the digital transformation literature, this study investigates how knowledge distance and political distance shape subsidiary exits through a U-shaped relationship, and how digital transformation breadth and depth differentially reconfigure these effects. We conduct empirical research on 1203 Chinese multinational enterprises from 2015 to 2019. The results indicate that both knowledge distance and political distance exhibit a U-shaped relationship with the subsidiary exit. The breadth of digital transformation strengthens the U-shaped relationship between knowledge distance and subsidiary exit but weakens the relationship between political distance and subsidiary exit. The depth of digital transformation mitigates the effects of both knowledge distance and political distance on subsidiary exit. These findings provide a novel explanatory perspective on the ‘Distance Paradox’ in internationalization theory, address a critical gap in the multinational enterprise (MNE) exit literature, and propose a modular governance blueprint for MNEs.

1. Introduction

Digital disruption and geopolitical fragmentation are converging. This convergence is fundamentally changing how multinational enterprises (MNEs) create and capture value across borders. In today’s era, knowledge is both a strategic asset and a geopolitical weapon [1]. MNEs face a paradoxical challenge: while digital technologies enable innovative business models, these technologies simultaneously redefine how cross-border distance impacts overseas operations. Meanwhile, against the backdrop of frequent geopolitical conflicts, the survival and exit decisions of MNEs’ subsidiaries are facing unprecedented institutional complexity [2]. For instance, the U.S.–China trade war and technological rivalry have heightened barriers to knowledge flows and policy restrictions, increasing operational costs and political risks for firms in each other’s markets. In such a complex environment, overseas subsidiaries face greater obstacles to knowledge transfer and heightened policy uncertainty. When tensions between countries escalate, subsidiaries may become casualties of geopolitical maneuvering, compelled to exit the market. Therefore, studying the impact of knowledge distance and political distance on subsidiary exit not only helps explain corporate behavior in the ongoing restructuring of global value chains but also provides a theoretical framework for MNEs to develop “technology–political” ambidextrous strategies in the digital age. The urgency of this issue lies in the fact that when knowledge becomes a tool of geopolitical competition, its distance effects escalate from an economic concern to a matter of survival.
The existing research on distance and subsidiary exit shows inconsistent findings [3,4,5], suggesting that distance factors may be moderated by other conditions. Thus, linear models may be inadequate. This study proposes that knowledge and political distance may nonlinearly influence subsidiary exit, requiring theoretical and empirical testing. Using institutional theory, we construct a “knowledge–political dual distance” framework with data from Chinese MNEs from 2015 to 2019. As global firms frequently enter and exit markets, understanding how these distances nonlinearly affect exit decisions is urgent. This enriches international business (IB) theory on distance and exit, responds to calls for more exit strategy research [6], and offers practical guidance for MNEs in volatile environments.
Digital transformation centralizes global business change, reshaping MNE strategies and operations. Technologies like AI, IoT, and blockchain bring new opportunities and challenges [7]. Against this backdrop, we introduce the breadth and depth of digital transformation as moderators to examine how they alter the distance–exit relationship. Current IB digital studies often focus on single dimensions (e.g., digital investment) [8], overlooking the dual aspects of breadth (coverage) and depth (penetration). As digital practices are multidimensional, studying how breadth and depth moderate nonlinear distance effects offers theoretical value. Our contingency perspective reveals under which digital transformation conditions knowledge and political distance exert the most pronounced nonlinear effects on MNEs’ subsidiary exit decisions, addressing the academic call for “examining the effects of digital transformation on developing economies” [7].
In summary, the current international landscape demands a re-evaluation of the risks and opportunities in MNE overseas operations. Geopolitical uncertainties highlight the need to examine how factors like knowledge distance and political distance affect subsidiary survival, while the wave of digital transformation prompts us to rethink how technological capabilities reshape traditional distance mechanisms.

2. Theoretical Development and Hypotheses

The extant IB literature has predominantly emphasized the impact of institutional, geographic, and cultural distances on MNEs’ strategic decision-making [9,10,11]. However, recent scholarship has increasingly highlighted the significance of knowledge distance and political distance [1,12]. Unlike traditional Western multinational corporations that expand overseas primarily to seek markets, a hallmark motivation for multinational enterprises from emerging markets like China is the “springboard perspective” [13]. This refers to leveraging international expansion as a means to rapidly acquire strategic assets (such as technology and knowledge) and achieve a competitive catch-up on a global scale. While a substantial knowledge gap is a primary driver for such investments, it may also become a critical source of operational failure and eventual exit. Simultaneously, the internationalization of Chinese enterprises is profoundly influenced by domestic industrial policies, national strategies (e.g., the Belt and Road Initiative), and the political environment of host countries. This implies that their investment decisions are often guided by political considerations rather than purely market-driven motives. Therefore, differences in political systems, ideologies, and international alliances between the host country and China can directly undermine the stability of investments. Thus, political distance constitutes a crucial yet frequently underestimated dimension affecting subsidiary survival. The two dimensions of knowledge distance and political distance, representing internal and external factors, respectively, form a strong framework for explaining the survival or exit of Chinese overseas subsidiaries. Grounded in institutional theory, which underscores how institutional environments shape organizational behavior and performance, this perspective posits that strategic decisions are not only driven by economic rationality but are also constrained by both informal institutions (cognitive frameworks) and formal institutions (legal and policy systems).
Current research recognizes that both knowledge distance and political distance influence subsidiary survival, yet their effects often yield inconsistent or even contradictory conclusions depending on the political and technological context. Empirical studies suggest that these distances can simultaneously generate positive and negative effects for multinational enterprises. For instance, knowledge distance may stimulate investment intentions yet hinder knowledge absorption and innovation performance; meanwhile, political distance, while amplifying compliance challenges, does not necessarily lead to subsidiary exit [14,15,16,17]. These contradictions imply that knowledge distance and political distance likely influence corporate decision-making through complex nonlinear mechanisms. Therefore, it is essential to further investigate their nonlinear effects and introduce corporate digital transformation as a key moderating variable to analyze how it alters the impact of these distances on divestment decisions.
Existing digital literature primarily investigates the direct effects of digital transformation on firm internationalization [7], noting that digital transformation helps firms acquire and disseminate knowledge across borders while presenting new challenges for international operations [18]. Furthermore, when examining the impact of knowledge distance and political distance on overseas subsidiary exit, existing studies mainly focus on host country institutional environments and cultural differences [3], overlooking the potential moderating role of firms’ digital transformation. We categorize digital transformation into two dimensions: breadth and depth, which, respectively, influence firms’ effective operations through information coverage scope and technological embeddedness [19]. Specifically, digital transformation breadth reflects the range of diverse digital technologies adopted by firms, while digital transformation depth reflects the degree of integration between digital technologies and firm operations [20]. Notably, as two differentiated strategic approaches, digital transformation breadth (coverage scope) and depth (degree of technological integration) may generate different contingent effects on the impact of knowledge distance and political distance. Therefore, this study proposes that digital capabilities can moderate the effect of knowledge and political distance on exit decisions, enabling firms to respond more flexibly to international challenges.
Building on institutional theory, we construct a “knowledge–political dual distance” framework and utilize data from 1203 Chinese multinational enterprises from 2015 to 2019 to examine the impact of knowledge and political distance on overseas subsidiary exit, while further investigating the moderating effects of digital transformation breadth and depth. This study aims to address two key research questions: (1) How do knowledge and political distance affect overseas subsidiary exit? (2) How do the breadth and depth of digital transformation moderate the relationship between knowledge and political distance and subsidiary exit? Through this framework, our research not only fills a gap in existing theory regarding knowledge–political distance and digital moderation mechanisms, but also provides new insights for multinational enterprises to make strategic adjustments in complex international environments.

2.1. The U-Shaped Effects of Knowledge Distance on Subsidiary Exit

In multinational operations, knowledge distance is considered to reflect the degree of technological, managerial, and market cognition differences between the parent company and its overseas subsidiaries, which affects the efficiency of knowledge transfer and strategic synergy between them [1,21]. A small knowledge distance implies a highly homogeneous market environment for the subsidiary [3]. Due to the ease of imitation in technology, products, and business models, the market becomes saturated with competitors, leading to compressed profit margins. The subsidiary struggles to establish a differentiated advantage, which may result in forced exit due to unsustainable profitability.
When knowledge distance falls within a moderate range, subsidiaries can leverage the parent company’s technological or managerial advantages to develop differentiated competitiveness in the host country. First, a moderate increase in knowledge distance helps subsidiaries from developing countries avoid excessive competition. As knowledge distance increases, differences in national innovation systems and technological capabilities widen, enabling multinational subsidiaries to leverage these gaps to establish unique technological advantages, create entry barriers, reduce direct competition, and improve survival rates [3,22]. For instance, Transsion Holdings Limited focuses on emerging markets, where its moderate technological leadership has allowed it to capture over 40% of the African market. Local competitors cannot replicate Transsion’s integrated hardware–software capabilities, while Western brands are unwilling to customize low-cost models for African consumers. Second, when knowledge distance remains within a moderate range, subsidiaries in developed countries gain more opportunities to acquire unique knowledge resources unavailable to the parent company [23], such as specialized production techniques or unique market demand insights. These distinctive knowledge assets provide competitive advantages, facilitating market expansion and reducing the likelihood of subsidiary exit.
However, when knowledge distance continues to increase, subsidiaries face escalating challenges in knowledge acquisition and integration, making knowledge exchange and alignment between the parent and subsidiary extremely difficult [24,25]. Subsidiaries must invest significantly more resources and effort to comprehend and apply the parent company’s knowledge while also localizing it to fit the host country’s context. Simultaneously, the parent company struggles to effectively transfer core knowledge to the subsidiary, and the subsidiary cannot efficiently feed locally acquired knowledge back to the parent for optimization. Such vast disparities in knowledge systems result in the parent company’s management practices failing to take root in the subsidiary, leading to operational inefficiencies [26]. Moreover, excessive knowledge distance creates coordination difficulties in decision-making and strategic planning [27]. Divergent perspectives on management, market trends, and business direction may arise, substantially increasing coordination costs and knowledge transfer expenses, further deteriorating subsidiary performance and prompting exit [15]. Subsidiaries may struggle to collaborate effectively with the parent company, leading to operational difficulties in the host market and increasing exit. To illustrate the challenge of knowledge distance in international business, consider the case of Best Buy, a U.S. retail giant, which faced significant difficulties in China. Best Buy’s mature supply chain management and membership-based model clashed with local consumer preferences (e.g., in-store bargaining) and the flexible strategies of domestic competitors (e.g., Gome, Suning). Unable to integrate local knowledge and adapt its business model, Best Buy sustained losses and ultimately exited the Chinese market (this is presented as an illustrative case). Therefore, we hypothesize the following statement:
H1. 
The impact of knowledge distance on subsidiary exit is U-shaped. Subsidiary exit decreases as knowledge distance increases from low to medium levels and increases as distance rises from medium to high levels.

2.2. The U-Shaped Effects of Political Distance on Subsidiary Exit

Political distance is considered to reflect the degree of differences between the home and host countries in terms of political systems, policies and regulations, and governance environments, which influences the strategic decision-making and risk management of multinational enterprises [16]. An excessively small political distance implies that their political systems, policies, and market rules are highly similar, resulting in relatively limited access to unique resources. This institutional homogeneity makes it difficult for multinational enterprises to establish distinctive competitive advantages. For example, when the U.S. retail giant Target entered the Canadian market in 2013, the high similarity in business regulations, consumer habits, and political–economic environments between the U.S. and Canada prevented Target from developing a differentiated edge. Meanwhile, local Canadian retailers and existing U.S. competitors dominated the market with their well established localized operations. The intense homogenized competition led to sustained losses for Target, ultimately forcing its complete exit from the Canadian market in 2015 (this is presented as an illustrative case).
When political distance falls within a moderate range, firms benefit from a relatively stable and predictable operating environment while also gaining opportunities for institutional arbitrage [28,29]. As political distance increases, overseas subsidiaries gain access to markets with distinct political environments, enabling them to acquire unique local resources such as specific policy incentives or strategic assets, thereby reducing exit [30,31]. Additionally, a greater political distance amplifies differences in market demand and consumer behavior between the host and home countries, creating a broader market space and new opportunities for subsidiaries [32,33]. By catering to local market needs, subsidiaries can expand their market share, enhance competitiveness, and mitigate exit risks.
However, when political distance becomes excessively large, overseas subsidiaries face significant political risks [14,17]. Specifically, pronounced differences in political systems may exacerbate conflicts between subsidiaries and the host country’s political environment by widening legal and cognitive distances [28]. This institutional friction not only reduces the subsidiary’s conflict resolution efficiency but may also lead to substantial increases in operational costs, such as those arising from weak intellectual property protection, ultimately elevating the risk of market exit. Furthermore, excessive political distance undermines the subsidiary’s ability to establish legitimacy in the host country due to issues like political instability and strained international relations [16,34]. For instance, disparities in political stability or geopolitical tensions may trigger policy volatility, surging compliance costs, or even political boycotts, making it difficult for firms to build sustainable legitimacy within the host country’s institutional environment [35]. This legitimacy deficit not only hampers trust-building with local stakeholders but may also lead to deteriorating operational efficiency due to regulatory conflicts or social exclusion, eventually forcing exit. A case in point is Google, which struggled to reconcile its global business model with China’s local regulatory requirements (e.g., content censorship) due to the large political distance (this is presented as an illustrative case). This compliance conflict ultimately led to Google’s exit from the mainland Chinese market. Therefore, we hypothesize the following statement:
H2. 
The impact of political distance on subsidiary exit is U-shaped.

2.3. The Moderation Effect of Breadth of Digital Transformation

Organizational information technology theory suggests that digital breadth reflects the diversity and complexity of a firm’s digital infrastructure [36,37]. When a parent company possesses high digital breadth, the heterogeneity of its knowledge management systems, data standards, and collaboration tools can impair the knowledge processing capabilities of overseas subsidiaries [20]. Consequently, in situations where knowledge distance is relatively low, extensive digital breadth leads to information redundancy. This means subsidiaries must simultaneously process duplicate or conflicting information from multiple digital channels (e.g., ERP, CRM, BI systems), increasing filtering costs and reducing operational efficiency [38]. This “information overload” effect undermines the knowledge resource advantages that low knowledge distance should provide, making subsidiaries more likely to exit due to management chaos [38,39].
Additionally, parent companies with high digital breadth often prefer standardized digital tools (e.g., SAP, Oracle), requiring subsidiaries to strictly adhere to their data architectures and processes [40]. Such standardization may inhibit a subsidiary’s ability to adjust its knowledge application and product strategies based on local market characteristics, making it difficult to balance “adaptation” and “standardization”, leading to “adaptation inflexibility” [41]. Complex digital interfaces and system conflicts can hinder the effective transfer of tacit knowledge, preventing the realization of optimal knowledge distance benefits.
When there is high knowledge distance between the parent company and subsidiary, the negative effects of the parent company’s extensive digital breadth (i.e., diverse digital systems and platforms) are further amplified, particularly in terms of technological conflicts and surging knowledge integration costs. Specifically, the parent company’s diverse digital systems (e.g., hybrid cloud architectures, multi-standard data protocols) may create compatibility issues with the subsidiary’s local systems, causing knowledge integration costs to rise exponentially [42]. For example, when a parent company simultaneously adopts ISO and GB data standards, subsidiaries may face additional burdens such as data cleansing and format conversion, delaying decision-making and increasing operational risks. Moreover, when digital breadth is high, subsidiaries may be forced to establish independent digital systems to reduce coordination costs, further widening the knowledge gap and creating “digital centrifugal forces”, ultimately accelerating the exit of overseas subsidiaries.
Overall, digital breadth not only amplifies the negative effects of knowledge distance but also makes the U-shaped relationship more pronounced. Therefore, we hypothesize the following statement:
H3. 
The breadth of digital transformation sharpens the U-shaped curve between knowledge distance and subsidiary exit.
Institutional theory posits that political distance reflects the differences between the home and host countries in terms of political systems, policy environments, and regulatory frameworks. When firms possess high digital breadth, their diversified digital infrastructure can mitigate the negative effects of political distance through the following mechanisms:
Under relatively low political distance, higher digital breadth enhances a firm’s institutional arbitrage capability. An extensive digital reach enables firms to more promptly and comprehensively access host country-specific policy incentives (e.g., tax reductions, R&D subsidies) and strategic resources (e.g., critical technology licenses, local high-quality suppliers) [43,44]. This improves resource acquisition efficiency, thereby strengthening subsidiary competitiveness and viability while reducing the likelihood of exit.
In cases of excessively large political distance, the information redundancy effect associated with digital breadth can transform into an early warning advantage. Broad digital breadth allows parent companies to leverage multi-source data monitoring systems to cross-validate and integrate political, economic, and social dynamics in the host country [45]. This fosters a more comprehensive understanding of the host country’s overall environment, reducing biases from single information sources. Through the continuous monitoring and advanced analysis of such multidimensional data, firms can detect subtle signals or anomalous patterns that foreshadow political risks earlier [46]. This enables the more accurate pre-emptive identification of potential policy changes, allowing subsidiaries to take proactive measures and minimizing exit risks stemming from political uncertainty. Therefore, we hypothesize the following statement:
H4. 
The breadth of digital transformation flattens the U-shaped curve between political distance and subsidiary exit.

2.4. The Moderation Effect of Depth of Digital Transformation

Under conditions of small knowledge distance, higher digital depth can further amplify the effects of knowledge synergy, enabling firms to utilize and share knowledge more effectively. Specifically, digital depth allows firms to leverage data analytics tools and other digital means to more thoroughly extract value from the knowledge resources acquired by subsidiaries in local markets [47]. This optimization of subsidiary business strategies enhances competitiveness and reinforces the advantages conferred by knowledge distance.
Moreover, greater digital depth facilitates the more efficient sharing of unique knowledge resources between parent companies and subsidiaries [48]. Parent companies can rapidly transfer management and operational expertise to subsidiaries through digital platforms, while subsidiaries can promptly feed back local market insights and other distinctive knowledge [49]. This high-precision knowledge sharing enables subsidiaries to better integrate global knowledge from headquarters with locally acquired insights, thereby more effectively developing local operations, improving operational efficiency and innovation capabilities, and further reducing the likelihood of involuntary exit from host markets.
Excessive knowledge distance is a key cause of knowledge transfer failures, particularly for highly contextualized, complex, or tacit knowledge [1]. Firms may struggle to accurately comprehend and effectively absorb external knowledge due to a lack of relevant background knowledge, cognitive frameworks, or practical experience [50]. Digital depth mitigates the risk of failure in knowledge transfer, absorption, and recreation processes caused by significant knowledge system disparities [51], thereby reducing the knowledge acquisition and integration challenges faced by subsidiaries. This substantially improves the success rate and effectiveness of transferring complex and tacit knowledge across large knowledge distances. For instance, Siemens’ Industry 4.0 knowledge base achieved 58% higher knowledge absorption efficiency in Chinese factories through its “AI-driven knowledge conversion system” (neuro-symbolic AI), which automatically adapts German engineering standards to local operational specifications (this is presented as an illustrative case). Therefore, we hypothesize the following statement:
H5. 
The depth of digital transformation flattens the U-shaped curve between knowledge distance and subsidiary exit.
Under conditions of relatively small political distance, MNEs not only need to access host country policy incentives and strategic resources but also must efficiently integrate these external resources with their internal capabilities to develop unique competitive advantages. A parent company’s high level of digital depth plays a crucial role in this process.
By leveraging advanced digital tools, such as customer relationship management systems with integrated analytics modules, firms can systematically optimize the entire process of identifying, evaluating, acquiring, and integrating host country policy incentives and strategic resources [18]. Furthermore, through digital management systems, parent companies can better coordinate subsidiaries in obtaining local-specific policy benefits and strategic assets [52]. At the same time, greater digital depth enables firms to employ sophisticated data analytics tools to assess the value of resources and facilitate their integration [53], thereby enhancing subsidiary competitiveness and further reducing the likelihood of exit.
When political distance is excessively large, the external environment is often characterized by uncertainty, potential risks, and barriers to resource acquisition [12]. In such contexts, MNEs and their subsidiaries must rely more heavily on internal capabilities, operational efficiency, and precise environmental insights to sustain survival and growth. Through big data analytics systems, parent companies can conduct the deep mining and pattern recognition of multidimensional market data from subsidiary locations, enabling firms to more accurately identify subtle policy shifts that traditional methods might overlook [41]. Additionally, digital depth enhances internal operational processes, improving subsidiary efficiency and reinforcing stable local operations, thereby reducing exit risks [54].
Moreover, when original target markets or critical resource supply channels are disrupted due to political factors, parent companies—equipped with digital depth (e.g., global supply chain visualization and risk analysis platforms, dynamically updated global market intelligence systems)—can help subsidiaries more effectively identify, evaluate, and swiftly transition to alternative market opportunities or resource supply channels worldwide. This capability mitigates risks and ensures operational continuity amid political turbulence and sudden policy changes [45].
In summary, the operational resilience and strategic flexibility conferred by high digital depth significantly strengthen a subsidiary’s survival capacity and long-term viability in adverse political environments, fundamentally reducing the risk of forced exit. Therefore, we hypothesize the following statement:
H6. 
The depth of digital transformation flattens the U-shaped curve between political distance and subsidiary exit.

3. Materials and Methods

3.1. Sample and Data Collection

We collected data from multiple sources including the CSMAR database, World Governance Indicators (WGIs), the Heritage Foundation, and the China Statistical Yearbook. To test our hypotheses, we compiled a sample of overseas subsidiaries of Chinese multinational corporations, focusing on companies listed on the Shanghai or Shenzhen Stock Exchanges between 2015 and 2019. Our final sample consists of 1203 overseas subsidiaries from Chinese multinational enterprises operating across 60 countries.
The sample firms have an average age of 20 years and an average total asset value of CNY 34.6 billion. These firms span 67 industries, with computer, communication, and other electronic equipment manufacturing comprising the largest number of enterprises (1722). A total of 3018 firms operate in high-digital-intensity sectors, such as telecommunications and information technology.
According to the classification standards of the International Monetary Fund (IMF), among the 60 countries listed, 26 are advanced economies and 34 are emerging and developing economies. The advanced economies include Australia, Austria, Belgium, Canada, Switzerland, the Czech Republic, Germany, Denmark, Spain, Estonia, Finland, France, the United Kingdom, Hungary, Italy, Japan, Korea, Lithuania, Latvia, the Netherlands, Norway, New Zealand, Poland, Slovakia, Sweden, and the United States. The emerging and developing economies comprise Albania, Argentina, Azerbaijan, Bangladesh, Bulgaria, Bahrain, Bosnia and Herzegovina, Belarus, Bolivia, Brazil, Chile, Colombia, Ecuador, Ethiopia, Guatemala, Indonesia, Iran, Jordan, Kazakhstan, Kyrgyzstan, Morocco, Mexico, Malaysia, Nigeria, Pakistan, Peru, the Philippines, Russia, Thailand, Tunisia, Türkiye, Ukraine, Vietnam, and Zimbabwe. Among them, the United States has the highest number of subsidiaries (3314), followed by Germany (1302), and Australia ranks third (1041).
This study applied the following screening criteria to refine the sample: (1) excluding subsidiaries located in “tax havens” such as the Cayman Islands, Bermuda, and the British Virgin Islands, as well as countries with missing data; (2) removing ST and PT listed companies with abnormal financial data; (3) excluding financial sector listed enterprises; (4) eliminating firms with critical missing financial data during their operational period. This rigorous selection process ensures the reliability and validity of our dataset for empirical analysis.

3.2. Measurements

3.2.1. Dependent Variable

Subsidiary exit (Exit). We identify exit events by observing whether a firm’s foreign subsidiary has not been recorded in the database for two or more consecutive years and excluding cases such as subsidiary renaming. A value of 0 indicates that the firm’s foreign operations are still observed, while 1 indicates that the firm has made an exit decision (i.e., the firm no longer operates in that country) [55].

3.2.2. Independent Variables

Knowledge distance (KD) and political distance (PD). We employ knowledge and political distance measures from Berry et al. (2010) (the Mahalanobis distance), where knowledge distance captures disparities in patent portfolios and scientific outputs, while political distance reflects differences in political stability, democratic institutions, and trade bloc affiliations [3]. The Mahalanobis distance is selected as the core metric due to its theoretical capacity to handle correlated high-dimensional variables and ensure empirical validity. In contrast, the Euclidean distance assumes variable independence, disregarding inherent covariance structures. This limitation artificially inflates the contribution of correlated features through redundant weighting, thereby distorting true cross-country differences. For example, technical interdependencies between patents and scientific outputs may compound measurement bias in knowledge distance.

3.2.3. Mediator Variables

Breadth of digital transformation (BDT) and depth of digital transformation (DDT). Breadth of digital transformation (BDT) refers to the number of distinct types of digital technologies adopted by a firm, reflecting the diversity of technological coverage. Based on the CSMAR classification of five technology categories (Artificial Intelligence Technology, Cloud Computing Technology, Blockchain Technology, Big Data Technology, Digital Technology Applications), a firm is considered to have adopted a specific technology type if keywords associated with that category are mentioned in its annual financial report or annual report. For example: Company A mentions “intelligent data analysis” and “deep learning” (Artificial Intelligence Technology) and “data mining” (Big Data Technology) in its annual financial report or annual report, resulting in BDT = 2. The sample mean of BDT is 1.9, with a standard deviation of 1.4 and a kurtosis of 1.9. Depth of digital transformation (DDT) refers to the aggregate application intensity of all digital technologies adopted by a firm, measured as the total frequency of mentions of the aforementioned five technology types disclosed in the public reports of listed companies. This reflects the depth of technology implementation. Frequency is defined as the count of keywords related to a specific technology category appearing in the annual report/announcement. For example: Company B mentions keywords related to Blockchain Technology four times and Big Data Technology two times, resulting in DDT = 6. The sample mean of DDT is 15.6, with a standard deviation of 33.0 and a kurtosis of 43.0.

3.2.4. Control Variables

Drawing on the prior literature, we include the following control variables: (1) firm size (Size): natural logarithm of total assets. (2) Total number of top management team (TMT): total number of directors, supervisors, and senior executives. (3) Management ownership (MO): ratio of shares held by management to total shares outstanding. (4) Profitability (ROE): return on equity, calculated as net profit divided by shareholders’ equity. (5) Financial leverage (FL): debt-to-asset ratio, calculated as total liabilities divided by total assets. (6) Government support (GS): total government subsidies received in the observation year. (7) Economic freedom (EF): host country’s economic freedom index. (8) Cultural distance (CD): differences in attitudes toward authority, trust, individuality, and importance of work and family.

3.3. Model

This study employs a quantitative, non-experimental, explanatory, and longitudinal design to address its research questions and test its hypotheses. The analysis is based on a linear probability model (LPM) with high-dimensional fixed effects. The primary analysis adopts a firm–host country–year panel data structure. The model incorporates firm fixed effects, host country fixed effects, and year fixed effects. These fixed effects are implemented using Stata’s reghdfe command with the absorb option for estimation. Aligning with the methods of Haans et al. (2016), we adopt a quadratic term specification as our baseline analytical approach [56].
To examine the impact of knowledge distance and political distance on subsidiary exit, we construct the following models:
Exitijt = α0 + σCONit + δi + εj + ϒt + μijt
Exitijt = α0 + α1KDijt + α2KD2ijt + σCONit + δi + εj + ϒt + μijt
Exitijt = α0 + α1PDijt + α2PD2ijt + σCONit + δi + εj + ϒt + μijt
Exitijt = α0 + α1KDijt + α2KD2ijt + α3BDTit + α4KDijt × BDTit + α5KD2ijt × BDTit + σCONit + δi + εj + ϒt + μijt
Exitijt = α0 + α1PDijt + α2PD2ijt + α3BDTit + α4PDijt × BDTit + α5PD2ijt × BDTit + σCONit + δi + εj + ϒt + μijt
Exitijt = α0 + α1KDijt + α2KD2ijt + α3DDTit + α4KDijt × DDTit + α5KD2ijt × DDTit + σCONit + δi + εj + ϒt + μijt
Exitijt = α0 + α1PDijt + α2PD2ijt + α3DDTit + α4PDijt × DDTit + α5PD2ijt × DDTit + σCONit + δi + εj + ϒt + μijt
where i represents the surveyed sample firm, j denotes the host country of the subsidiary, and t indicates the observation year. Exit represents the exit decision of subsidiaries in each observation year; KD and PD denote knowledge distance and political distance, respectively; BDT and DDT represent digital transformation breadth and depth, respectively; CON includes all control variables; δi denotes firm fixed effects; εj denotes host country fixed effects; ϒt denotes year fixed effects; α0 is the intercept term; and μ is the error term.
Models (2) and (3) examine the direct effects of knowledge distance and political distance on overseas subsidiary exit. Models (4) and (5) test the moderating effect of BDT. Models (6) and (7) assess the moderating effect of DDT.
Prior to regression analysis, we conducted a descriptive statistics and pairwise correlation analysis using STATA(16.0). As summarized in Table 1, the correlation coefficients between variables were modest in magnitude and the standard deviations were low, indicating stable data properties suitable for regression analysis. Figure 1 further visualizes these relationships through a correlation heatmap, with Pearson’s correlation coefficients displayed in each cell. Additionally, variance inflation factor (VIF) diagnostics confirmed the absence of significant multicollinearity concerns, with all VIF values substantially below the threshold of 10.
Figure 2 shows the descriptive distributions (density) of KD, PD, BDT, and DDT, with all variables standardized to a mean of 0 and a standard deviation of 1.

4. Results

4.1. Hypothesis Testing Results

As shown in Model 2 of Table 2, the squared term of knowledge distance has a significant positive impact on subsidiary exit (b = 0.409, p < 0.05). The Sasabuchi test results show a p-value of less than 0.05, significantly rejecting the null hypothesis that no extreme point exists. The turning point is calculated at 10.19 on the mean-centered scale, with a 95% confidence interval of [1.51, 18.86]. After retransforming to the original scale, the optimal knowledge distance is 23.23, which is at the 69th percentile of the empirical distribution. This means that as knowledge distance moves from moderate to high levels, the likelihood of a subsidiary exiting increases sharply, which aligns with the U-shaped pattern proposed in H1; for example, companies with large technological differences face higher integration costs that drive them to withdraw from markets.
Model 3 indicates that the squared term of political distance also has a significant positive effect on subsidiary exit (b = 0.110, p < 0.01). The Sasabuchi test results show a p-value of less than 0.05, significantly rejecting the null hypothesis that no extreme point exists. The turning point is calculated at −2.52 on the mean-centered scale, with a 95% confidence interval of [−8.31, 3.28]. After retransforming to the original scale, the optimal knowledge distance is 12.44, which is at the 38th percentile of the empirical distribution. This implies that as political distance increases from low to moderate levels, subsidiary exit decreases, whereas when political distance moves from moderate to high levels, subsidiary exit increases. This pattern aligns with the U-shaped relationship proposed in H2.
In Models 4 and 5, we find that the interaction term between BDT and the squared term of KD has a significantly positive effect on overseas subsidiary exit (b = 0.072, p < 0.01), while the interaction term between BDT and the squared term of PD has a significantly negative effect (b = −0.081, p < 0.01). Specifically, the positive coefficient (0.072) indicates that BDT amplifies the U-shaped relationship between knowledge distance and overseas subsidiary exit, whereas the negative coefficient (−0.081) suggests that BDT mitigates the U-shaped relationship between political distance and exit behavior. These findings provide strong quantitative support for H3 and H4.
In Models 6 and 7, we observe that the interaction term between DDT and the squared term of KD has a significantly negative impact on overseas subsidiary exit (b = −0.051, p < 0.01), and the interaction term between DDT and the squared term of PD also has a significantly negative effect (b = −0.052, p < 0.01). The significantly negative coefficients quantitatively demonstrate that DDT mitigates the uncertainties and risks associated with both knowledge and political distances, thereby reducing subsidiary exit at both the high and low levels of these distances, supporting H5 and H6.

4.2. Heterogeneity Analysis

To further elucidate the mechanisms through which digital breadth and depth operate, this study categorizes the sample into developed economies and developing economies. As shown in Table 3, in the subsample of developed economies, BDT continues to exhibit a significant exacerbating effect on the U-shaped curve of KD (b = 0.098, p < 0.01), while it significantly flattens the U-shaped curve of PD (b = −0.061, p < 0.01). In the subsample of developing economies, the aggravating effect of BDT on the U-shaped relationship for KD becomes statistically insignificant; however, its mitigating effect on the U-shaped curve of PD remains significant (b = −0.081, p < 0.05). Meanwhile, DDT demonstrates a significant flattening effect on the U-shaped curves of both KD and PD across both groups.

4.3. Robustness Check

First, we conduct robustness tests by replacing the baseline regression model with the Cox model to re-estimate the main effects. As shown in Table 4, the key conclusions remain valid under the Cox model, further confirming the robustness of our findings.
Second, we employ industry−year-standardized data to measure BDT and DDT. As shown in Table 5, the core findings are consistent with the preliminary regression findings, verifying the robustness of the model.
Finally, considering that industries such as telecommunications, information technology, and related sectors typically possess a stronger digital foundation and more advanced internal digital transformation processes, we exclude high-digitalization industries, including telecommunications, radio and television satellite transmission services, software and IT services, and internet-related services, from the regression analysis [57]. The regression results in Table 6 demonstrate that the research conclusions remain robust even after excluding these specific industries.

4.4. Endogeneity Test

We use two methods to address potential endogeneity issues.
First, we use Propensity Score Matching (PSM) to address potential self-selection bias in knowledge distance and political distance. We employed the Inverse Probability Weighting (IPW) method to handle the continuous distance variable. After matching, all variables exhibited a Standardized Mean Difference (SMD) of less than 0.1, satisfying the balance requirement for causal inference (Table 7). The regression results in Table 8 indicate that both knowledge distance and political distance have positive effects on overseas subsidiary exit.
Second, we use the nonlinear control function method (2SRI). Following the synthetic instrumental variable method of Bahar and Rapoport (2018) [58], we selected time-invariant external characteristics—namely enterprise ownership type, industry, and province—and generated interaction terms with time trends. The fitted values from these interactions serve as instrumental variables (IV). This design satisfies the following factors: ① relevance: the IV captures systematic variation in the explanatory variables driven by exogenous characteristics; ② exogeneity: the time-invariant characteristics are not subject to reverse causality from the explanatory variables and remain orthogonal to the error term. The diagnostic results show that the F statistic (90.69) far exceeds the Stock−Yogo 10% bias critical value, which rejects the weak instrument hypothesis. Since the number of instruments equals the number of endogenous variables (exactly identified), no overidentification issue exists. As presented in Table 9, both the residual term and its squared term are insignificant, suggesting the absence of endogeneity concerns.

5. Discussion

5.1. Theoretical Implications

First, this study provides a novel perspective on the ‘Distance Paradox’ in internationalization theory by revealing the nonlinear effects of knowledge distance and political distance on subsidiary exit [59]. While traditional institutional theory emphasizes that cross-border distances increase the liability of foreignness for multinational enterprises, existing research has inadequately explored the nonlinear effects of knowledge and political distances [14]. Breaking free from linear assumptions, this study uncovers a U-shaped relationship between knowledge/political distance and subsidiary exit—where both excessively low and high distances elevate exit risks, while moderate distances may stimulate organizational learning through “optimal conflict.” This finding not only challenges the linear paradigm of traditional distance theory but also offers a theoretical foundation for how MNEs can strategically manage distance effects to achieve international resilience.
Second, although digital technologies are regarded as effective tools for multinational enterprises to optimize global innovation resources and accelerate internationalization [60], scholars argue that MNEs may struggle to leverage digitalization for sustainable competitive advantages—due to insufficient institutional resources for digital innovation, limited opportunities to adopt and develop digital technologies, and difficulties in integrating digitalization into internationalization processes [61]. Our findings, however, demonstrate that MNEs can effectively utilize digital technologies to mitigate the impact of knowledge and political distances on subsidiary exit. Moreover, while existing studies predominantly focus on single technological dimensions (e.g., digital tool adoption) in internationalization [53], they overlook the differential effects of digital transformation breadth (technology coverage) and depth (technology embeddedness). Our research highlights the breadth and depth of digital transformation as key factors for MNEs navigating complex international environments, emphasizing the need for firms to reassess the weighting of distance factors in the digital era. By incorporating heterogeneity analysis based on host country development levels, our study provides clearer insights into the specific contextual conditions and boundary mechanisms through which the breadth and depth of digital technologies either mitigate or exacerbate the ‘Distance Paradox’ in multinational investment. This approach significantly deepens the theoretical understanding of how digital technologies reshape spatial relationships in international business operations. These findings not only enrich the proposition by Berry et al. (2010) regarding the impact of political distance on firms’ internationalization strategies but also broaden its scope by revealing that the depth of digital transformation can mitigate this effect in politically volatile environments, thereby providing novel insights into previously unexplored strategic pathways [3]. By addressing the call for “examining the effects of digital transformation on developing economies” [7], this study not only contributes to the discourse on digitalization but also provides a framework to bridge the contextual gap in MNE exit theory [57].

5.2. Managerial Implications

First, multinational corporations should consider the knowledge and political system differences between home and host countries when evaluating subsidiary retention or exit, as these differences entail nonlinear, complex risks. As exemplified by the case of Best Buy, a substantial knowledge gap existed between its U.S. retail system and the institutional logic governing China’s market. The measured knowledge distance (KD) significantly exceeded the empirically optimal threshold at the 69th percentile (reaching beyond the 95th percentile). Such pronounced divergence surpassed the organization’s absorptive capacity and adaptive capabilities, ultimately resulting in its market exit due to an irreversible failure to internalize and respond to contextual discrepancies. Furthermore, as exemplified by the previously discussed Target case, the political environment similarity between the United States and Canada faced by Target was excessively high (with its PD value falling below the fifth percentile). By entering a highly saturated market, the firm’s minimal institutional divergence proved insufficient to establish sustainable entry barriers or distinctive competitive advantages, ultimately resulting in its market exit. Moreover, these cases not only reflect the outcomes of corporate strategic choices but also reveal how multinational investment and exit actions exert profound ripple effects on broader business ecosystems. The exit of a subsidiary often disrupts local supply chains and dismantles knowledge-sharing networks, thereby reshaping regional market competition dynamics. Therefore, when formulating investment and operational strategies, multinational corporations must urgently incorporate these systemic ecological impacts. By identifying risks and opportunities through multidimensional distance management, they can achieve a dynamic balance between global strategy and local embeddedness.
Second, this study highlights the strategic importance of digital transformation in international operations, demonstrating how digital capabilities interact with institutional environments. Enhancing both the breadth and depth of digital transformation improves firms’ adaptability to diverse cultural and institutional settings, thereby increasing subsidiary survival likelihood. Specifically, BDT facilitates embedding within multi-market ecosystems and enables inter-firm collaboration, while DDT supports integrated yet flexible global value architectures that foster cross-border co-creation and systemic resilience. However, the implementation of digital strategies is constrained by the firm’s existing level of digital maturity and resource availability. Thus, although tailored digital strategies offer clear theoretical advantages, in practice they must be grounded in the firm’s own digital foundation and accessible resources. This aligns with multiple research findings on digital maturity. For instance, Krulčić et al. (2025) indicate that a firm’s digital maturity directly influences its technology adoption path and transformation outcomes [62]; similarly, Pörtner et al. (2025) emphasize that many organizations continue to face challenges in fully realizing the benefits of data, despite substantial digital investments, largely due to the absence of a comprehensive and validated maturity assessment method to guide the transformation pathway [63].
Specifically, multinational enterprises should avoid uniform digitalization models. Instead, parent companies ought to align contextual factors—such as subsidiary-specific conditions—with their own digital maturity to guide phased and differentiated digital deployment. Firms with high maturity and abundant resources can pursue standardized, integrated systems, while those with lower maturity or constrained resources should adopt modular, incremental approaches focused on core processes to mitigate risks. Such tailored strategies not only improve subsidiary performance but also facilitate interconnected value systems beyond organizational boundaries. By aligning digital breadth with ecosystem partnerships and deepening integration in core activities, firms can strengthen their roles as central nodes in global digital ecosystems.
Furthermore, enterprises should customize digital implementation based on knowledge and political distance, market maturity, strategic importance, and local institutions. From an ecosystem perspective, this involves designing digital platforms that enable multilateral interaction and value exchange among various participants, rather than solely focusing on internal operational efficiency. It is essential for enterprises to establish a flexible and scalable digital architecture, incorporating open interfaces and interoperability standards to reduce the cost and complexity of subsequent system integration. Digital platforms should also facilitate local knowledge accumulation and global dissemination to enhance absorptive capacity and contextual adaptability. This approach positions multinationals as ecosystem orchestrators, enabling cross-border co-creation with partners. Finally, integrated digital systems for policy monitoring, risk warning, and emergency response can mitigate institutional risks and turn political distance into competitive advantages.

5.3. Limitations and Future Directions

This study focuses on Chinese MNEs; therefore, the external validity of its findings is primarily limited to emerging market environments sharing similar institutional backgrounds. It is important to highlight that the internationalization strategies and performance of Chinese MNEs are significantly influenced by unique contextual factors, such as a strong policy-driven nature (e.g., the ‘Belt and Road Initiative’) and a high prevalence of state ownership. In contrast, MNEs from non-emerging markets, particularly those in Organisation for Economic Co-operation and Development (OECD) countries, are typically driven more by market-driven mechanisms. Consequently, the direct transferability of this study’s findings to the context of OECD country MNEs is limited. Future research could focus on examining the applicability of this study’s core findings within the context of OECD country MNEs using a similar framework or conduct systematic cross-context comparative analyses. Such efforts would deepen our understanding of the commonalities and differences in MNE behavior across diverse institutional settings and market development stages, thereby extending the external validity of relevant theories.
Second, future research could use more granular industry-specific data to explore how the digital ecosystem (e.g., the development of industry platforms, the degree of supply chain digitalization) facilitates or hinders the internationalization process of firms, particularly in traditional low-tech industries.
Third, future research could explore more granular country-level characteristics (such as specific institutional differences and bilateral relationships), employ finer-grained cultural dimension analyses, or integrate qualitative methods to gain deeper insights into exit mechanisms within specific regional or cultural contexts.
Finally, this study focuses exclusively on two dimensions of cross-national distance (KD and PD). This limited scope may affect the generalizability of our findings. Although KD and PD are particularly salient in the context of Chinese multinational enterprises, other dimensions, such as economic or geographic distance, may also influence internationalization strategies. Consequently, the conclusions drawn here may not be fully applicable in contexts where other dimensions constitute the primary source of distance. Nevertheless, this focused approach does not substantially undermine the robustness of our findings within the specific Chinese business environment. Our theoretical framework offers an in-depth investigation into the mechanisms through which KD and PD operate. Given the strong policy orientation of China’s outward foreign direct investment and its strategic pursuit of knowledge-intensive assets, these two dimensions are especially relevant. Thus, while the model may not be comprehensive from a broader theoretical perspective, it remains robust and highly pertinent for explaining the core phenomena under investigation. Future research could incorporate additional dimensions of distance to develop a more holistic model and test the boundary conditions of this study’s conclusions.

Author Contributions

Conceptualization, Z.Z. and L.W.; methodology, Z.Z.; software, Z.Z.; validation, Z.Z.; formal analysis, Z.Z.; resources, L.W.; data curation, Z.Z.; writing—original draft preparation, Z.Z.; writing—review and editing, L.W.; supervision, L.W.; funding acquisition, L.W. All authors have read and agreed to the published version of the manuscript.

Funding

National Natural Science Foundation of China: No.72374041; Humanities and Social Sciences Project of the Ministry of Education: 22YJAGJW006; The Soft Science Project of Shanghai Science and Technology Commission: No.25692103800.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The authors declare that the data on financial characteristics, governance structure, and government subsidies were taken from the China Stock Market and Accounting Research databases: https://data.csmar.com/ (accessed on 12 May 2025). Cross-border distance data was sourced from Berry, Guillén, and Zhou (2010): https://mgmt.wharton.upenn.edu/distance−data−downloads−guillen/ (accessed on 12 May 2025).

Conflicts of Interest

The authors declare no conflicts of interest.

Abbreviations

The following abbreviations are used in this manuscript:
SizeFirm Size
TMTTotal Number of Top Management Team
MOManagement Ownership
ROEReturn on Equity
FLFinancial Leverage
GSGovernment Subsidies
EFEconomic Freedom
CDCulture Distance
ExitOversea Subsidiary Exit
KDKnowledge Distance
PDPolitical Distance
BDTBreadth of Digital Transformation
DDTDepth of Digital Transformation
MNEsMultinational Enterprises
CSMARChina Stock Market and Accounting Research
OECDOrganisation for Economic Co-operation and Development
LPMLinear Probability Model
IPWInverse Probability Weighting
SMDStandardized Mean Difference
2SRITwo-Stage Residual Inclusion

References

  1. Ouyang, H.; Park, C.; Oh, C.H. The Impact of Reverse Knowledge Transfers on Exploitative and Exploratory Innovations in MNCs: The Role of Knowledge Distance. J. Bus. Res. 2024, 179, 114682. [Google Scholar] [CrossRef]
  2. Witt, M.A. De-Globalization: Theories, Predictions, and Opportunities for International Business Research. J. Int. Bus. Stud. 2019, 50, 1053–1077. [Google Scholar] [CrossRef]
  3. Berry, H.; Guillén, M.F.; Zhou, N. An Institutional Approach to Cross-National Distance. J. Int. Bus. Stud. 2010, 41, 1460–1480. [Google Scholar] [CrossRef]
  4. Kang, J.; Lee, J.Y.; Ghauri, P.N. The Interplay of Mahalanobis Distance and Firm Capabilities on MNC Subsidiary Exits from Host Countries. Manag. Int. Rev. 2017, 57, 379–409. [Google Scholar] [CrossRef]
  5. Lee, E.T.; Park, Y.-R.; Kwak, J. Knowledge Distance and Innovation Performance: The Moderating Role of Internationalization Breadth and Depth. Asian Bus. Manag. 2023, 22, 1131–1154. [Google Scholar] [CrossRef]
  6. Tan, Q.; Sousa, C.M.P. Giving a Fish or Teaching to Fish? Exploring the Effects of Home-Country Governmental Support on Foreign Exit Decisions. Int. Mark. Rev. 2020, 37, 1181–1203. [Google Scholar] [CrossRef]
  7. Feliciano-Cestero, M.M.; Ameen, N.; Kotabe, M.; Paul, J.; Signoret, M. Is Digital Transformation Threatened? A Systematic Literature Review of the Factors Influencing Firms’ Digital Transformation and Internationalization. J. Bus. Res. 2023, 157, 113546. [Google Scholar] [CrossRef]
  8. Nwankpa, J.K.; Roumani, Y.; Datta, P. Process Innovation in the Digital Age of Business: The Role of Digital Business Intensity and Knowledge Management. J. Knowl. Manag. 2022, 26, 1319–1341. [Google Scholar] [CrossRef]
  9. Aguilera-Caracuel, J.; Aragón-Correa, J.A.; Hurtado-Torres, N.E.; Rugman, A.M. The Effects of Institutional Distance and Headquarters’ Financial Performance on the Generation of Environmental Standards in Multinational Companies. J. Bus. Ethics 2012, 105, 461–474. [Google Scholar] [CrossRef]
  10. Ojala, A. Geographic, Cultural, and Psychic Distance to Foreign Markets in the Context of Small and New Ventures. Int. Bus. Rev. 2015, 24, 825–835. [Google Scholar] [CrossRef]
  11. Keig, D.L.; Brouthers, L.E.; Marshall, V.B. The Impact of Formal and Informal Institutional Distances on MNE Corporate Social Performance. Int. Bus. Rev. 2019, 28, 101584. [Google Scholar] [CrossRef]
  12. Cui, V.; Tian, X.; Yu, R. Fragmentation in International Business: A Geographic-and-Political Perspective. Int. Bus. Rev. 2025, 34, 102458. [Google Scholar] [CrossRef]
  13. Luo, Y.; Tung, R.L. International Expansion of Emerging Market Enterprises: A Springboard Perspective. J. Int. Bus. Stud. 2007, 38, 481–498. [Google Scholar] [CrossRef]
  14. Brown, L.W.; Liou, R.-S.; Hasija, D. Emerging Market Multinational Corporations’ Cross-Border Mergers and Acquisitions and Political Distance: Does Corporate Political Activity Matter? Thunderbird Int. Bus. Rev. 2023, 65, 533–546. [Google Scholar] [CrossRef]
  15. Lissillour, R.; Cui, Y.; Guesmi, K.; Chen, W.; Chen, Q. Value Network and Firm Performance: The Role of Knowledge Distance and Environmental Uncertainty. J. Knowl. Manag. 2023, 28, 44–68. [Google Scholar] [CrossRef]
  16. White, G.O.; Boddewyn, J.J.; Rajwani, T.; Hemphill, T.A. Regulator Vulnerabilities to Political Pressures and Political Tie Intensity: The Moderating Effects of Regulatory and Political Distance. Manag. Int. Rev. 2018, 58, 743–769. [Google Scholar] [CrossRef]
  17. Gonchar, K.; Greve, M. The Impact of Political Risk on FDI Exit Decisions. Econ. Syst. 2022, 46, 100975. [Google Scholar] [CrossRef]
  18. Nambisan, S.; Zahra, S.A.; Luo, Y. Global Platforms and Ecosystems: Implications for International Business Theories. J. Int. Bus. Stud. 2019, 50, 1464–1486. [Google Scholar] [CrossRef]
  19. Yin, W. Identifying the Pathways through Digital Transformation to Achieve Supply Chain Resilience: An fsQCA Approach. Environ. Sci. Pollut. Res. 2023, 30, 10867–10879. [Google Scholar] [CrossRef]
  20. Yang, Z.; Hu, W.; Shao, J.; Shou, Y.; He, Q. How Does Digitalization Alter the Paradox of Supply Base Concentration? The Effects of Digitalization Intensity and Breadth. Int. J. Oper. Prod. Manag. 2023, 43, 1690–1722. [Google Scholar] [CrossRef]
  21. Xia, T.; Liu, X. Cross-National Knowledge Distance and Host Country Innovation—The Mediating Roles of Local Talents and Third-Country Exporting. Thunderbird Int. Bus. Rev. 2025, 67, 99–113. [Google Scholar] [CrossRef]
  22. Hölzl, W.; Janger, J. Distance to the Frontier and the Perception of Innovation Barriers across European Countries. Res. Policy 2014, 43, 707–725. [Google Scholar] [CrossRef]
  23. Castellani, D.; Montresor, S.; Schubert, T.; Vezzani, A. Multinationality, R&D and Productivity: Evidence from the Top R&D Investors Worldwide. Int. Bus. Rev. 2017, 26, 405–416. [Google Scholar] [CrossRef]
  24. Capaldo, A.; Lavie, D.; Messeni Petruzzelli, A. Knowledge Maturity and the Scientific Value of Innovations: The Roles of Knowledge Distance and Adoption. J. Manag. 2017, 43, 503–533. [Google Scholar] [CrossRef]
  25. Lane, J.N.; Ganguli, I.; Gaule, P.; Guinan, E.; Lakhani, K.R. Engineering Serendipity: When Does Knowledge Sharing Lead to Knowledge Production? Strateg. Manag. J. 2021, 42, 1215–1244. [Google Scholar] [CrossRef]
  26. Wen, J.; Qualls, W.J.; Zeng, D. To Explore or Exploit: The Influence of Inter-Firm R&D Network Diversity and Structural Holes on Innovation Outcomes. Technovation 2021, 100, 102178. [Google Scholar] [CrossRef]
  27. Bryan Jean, R.-J.; Sinkovics, R.R.; Kim, D. Antecedents and Outcomes of Supplier Innovativeness in International Customer–Supplier Relationships: The Role of Knowledge Distance. Manag. Int. Rev. 2017, 57, 121–151. [Google Scholar]
  28. Fainshmidt, S.; White, G.O.; Cangioni, C. Legal Distance, Cognitive Distance, and Conflict Resolution in International Business Intellectual Property Disputes. J. Int. Manag. 2014, 20, 188–200. [Google Scholar] [CrossRef]
  29. Konara, P.; Shirodkar, V. Regulatory Institutional Distance and MNCs’ Subsidiary Performance: Climbing up vs. Climbing down the Institutional Ladder. J. Int. Manag. 2018, 24, 333–347. [Google Scholar] [CrossRef]
  30. Klopf, P.; Nell, P.C. How “Space” and “Place” Influence Subsidiary Host Country Political Embeddedness. Int. Bus. Rev. 2018, 27, 186–197. [Google Scholar] [CrossRef]
  31. Nguyen, H.T.T.; Larimo, J.; Ghauri, P. Understanding Foreign Divestment: The Impacts of Economic and Political Friction. J. Bus. Res. 2022, 139, 675–691. [Google Scholar] [CrossRef]
  32. Song, S. Divesting or Keeping Overseas Subsidiary Production under Rising Production Costs and Uncertain Market Demands in Host Countries. Int. Bus. Rev. 2024, 33, 102320. [Google Scholar] [CrossRef]
  33. Wu, X.; Xiong, Y. Regional Differences in Urban Residents’ Consumption Behaviour in China: From the Perspective of the Habit Formation Time Effect. Econ. Res.-Ekon. Istraživanja 2022, 36, 2117229. [Google Scholar] [CrossRef]
  34. Chao, M.C.-H.; Kumar, V. The Impact of Institutional Distance on the International Diversity–Performance Relationship. J. World Bus. 2010, 45, 93–103. [Google Scholar] [CrossRef]
  35. Adarkwah, G.K.; Dorobantu, S.; Sabe, C.A. Geopolitical Volatility and Subsidiary Investments. Strateg. Manag. J. 2024, 45, 2275–2306. [Google Scholar] [CrossRef]
  36. Nadkarni, S.; Prügl, R. Digital Transformation: A Review, Synthesis and Opportunities for Future Research. Manag. Rev. Q. 2021, 71, 233–341. [Google Scholar] [CrossRef]
  37. Kraus, S.; Durst, S.; Ferreira, J.J.; Veiga, P.; Kailer, N.; Weinmann, A. Digital Transformation in Business and Management Research: An Overview of the Current Status Quo. Int. J. Inf. Manag. 2022, 63, 102466. [Google Scholar] [CrossRef]
  38. Shahrzadi, L.; Mansouri, A.; Alavi, M.; Shabani, A. Causes, Consequences, and Strategies to Deal with Information Overload: A Scoping Review. Int. J. Inf. Manag. Data Insights 2024, 4, 100261. [Google Scholar] [CrossRef]
  39. Yu, H.; Fletcher, M.; Buck, T. Managing Digital Transformation during Re-Internationalization: Trajectories and Implications for Performance. J. Int. Manag. 2022, 28, 100947. [Google Scholar] [CrossRef]
  40. Banalieva, E.R.; Dhanaraj, C. Internalization Theory for the Digital Economy. J. Int. Bus. Stud. 2019, 50, 1372–1387. [Google Scholar] [CrossRef]
  41. Meyer, K.E.; Li, J.; Brouthers, K.D.; Jean, R.-J. International Business in the Digital Age: Global Strategies in a World of National Institutions. J. Int. Bus. Stud. 2023, 54, 577–598. [Google Scholar] [CrossRef]
  42. Rong, K.; Ling, Y.; Yang, T.; Huang, C. Cross-Border Data Transfer: Patterns and Discrepancies. J. Int. Bus. Policy 2025, 8, 10–32. [Google Scholar] [CrossRef]
  43. Pollman, E. Tech, Regulatory Arbitrage, and Limits. Eur. Bus. Organ. Law Rev. 2019, 20, 567–590. [Google Scholar] [CrossRef]
  44. Su, H.; Cai, F.; Huang, Y. Institutional Constraints and Exporting of Emerging-Market Firms: The Moderating Role of Innovation Capabilities and Digital Transformation. Manag. Decis. Econ. 2022, 43, 2641–2656. [Google Scholar] [CrossRef]
  45. Luo, Y. A General Framework of Digitization Risks in International Business. J. Int. Bus. Stud. 2022, 53, 344–361. [Google Scholar] [CrossRef]
  46. Jiang, K.; Chen, L.; Li, J.; Du, X. The Risk Effects of Corporate Digitalization: Exacerbate or Mitigate? Humanit. Soc. Sci. Commun. 2025, 12, 317. [Google Scholar] [CrossRef]
  47. Bertello, A.; Ferraris, A.; Bresciani, S.; Bernardi, P.D. Big Data Analytics (BDA) and Degree of Internationalization: The Interplay between Governance of BDA Infrastructure and BDA Capabilities. J. Manag. Gov. 2021, 25, 1035–1055. [Google Scholar] [CrossRef]
  48. Massa, S.; Annosi, M.C.; Marchegiani, L.; Messeni Petruzzelli, A. Digital Technologies and Knowledge Processes: New Emerging Strategies in International Business. J. Knowl. Manag. 2023, 27, 330–387. [Google Scholar] [CrossRef]
  49. Ira Putri Hutasoit, D.; Elisabeth, D.; Indra Sensuse, D. Knowledge Sharing for Multinational Corporations Using Technology: A Systematic Literature Review. In Proceedings of the 2020 3rd International Conference on Algorithms, Computing and Artificial Intelligence, Sanya, China, 24–26 December 2020; Association for Computing Machinery: New York, NY, USA, 2020; pp. 1–8. [Google Scholar]
  50. Camisón, C.; Forés, B. Knowledge Absorptive Capacity: New Insights for Its Conceptualization and Measurement. J. Bus. Res. 2010, 63, 707–715. [Google Scholar] [CrossRef]
  51. Bucher, J.; Bader, B.; Deller, J. Cross-Border Knowledge Transfer in the Digital Age: The Final Curtain Call for Long-Term International Assignments? J. Manag. Stud. 2024, 61, 1792–1824. [Google Scholar] [CrossRef]
  52. Teece, D.J. The Multinational Enterprise, Capabilities, and Digitalization: Governance and Growth with World Disorder. J. Int. Bus. Stud. 2025, 56, 7–22. [Google Scholar] [CrossRef]
  53. Autio, E.; Nambisan, S.; Thomas, L.D.W.; Wright, M. Digital Affordances, Spatial Affordances, and the Genesis of Entrepreneurial Ecosystems. Strateg. Entrep. J. 2018, 12, 72–95. [Google Scholar] [CrossRef]
  54. Lin, B.; Xie, Y. Does Digital Transformation Improve the Operational Efficiency of Chinese Power Enterprises? Util. Policy 2023, 82, 101542. [Google Scholar] [CrossRef]
  55. Tang, R.W. Institutional Unpredictability and Foreign Exit−reentry Dynamics: The Moderating Role of Foreign Ownership. J. World Bus. 2023, 58, 101389. [Google Scholar] [CrossRef]
  56. Haans, R.F.J.; Pieters, C.; He, Z. Thinking about U: Theorizing and Testing U—And Inverted U-shaped Relationships in Strategy Research. Strateg. Manag. J. 2016, 37, 1177–1195. [Google Scholar] [CrossRef]
  57. Huang, X.; Gao, Y. Integration of Digital and Real Industry Technologies and Enterprise Total Factor Productivity: A Study Based on Chinese Enterprise Patent Information. China Ind. Econ. 2023, 11, 118–136. [Google Scholar] [CrossRef]
  58. Bahar, D.; Rapoport, H. Migration, Knowledge Diffusion and the Comparative Advantage of Nations. Econ. J. 2018, 128, F273–F305. [Google Scholar] [CrossRef]
  59. Hutzschenreuter, T.; Kleindienst, I.; Lange, S. The Concept of Distance in International Business Research: A Review and Research Agenda. Int. J. Manag. Rev. 2016, 18, 160–179. [Google Scholar] [CrossRef]
  60. Deng, Z.; Zhu, Z.; Johanson, M.; Hilmersson, M. Rapid Internationalization and Exit of Exporters: The Role of Digital Platforms. Int. Bus. Rev. 2022, 31, 101896. [Google Scholar] [CrossRef]
  61. Li, W.; Wang, S.; Deng, X. The Impact of Digital Finance on Business Environment: Mediating Role of Industrial Structural Upgrading and Moderating Role of Digital Infrastructure. Financ. Res. Lett. 2024, 67, 105775. [Google Scholar] [CrossRef]
  62. Krulčić, E.; Doboviček, S.; Pavletić, D.; Čabrijan, I. A Dynamic Assessment of Digital Maturity in Industrial SMEs: An Adaptive AHP-Based Digital Maturity Model (DMM)with Customizable Weighting and Multidimensional Classification (DAMA-AHP). Technologies 2025, 13, 282. [Google Scholar] [CrossRef]
  63. Pörtner, L.; Riel, A.; Schmidt, B.; Leclaire, M.; Möske, R. Data Management Maturity Model—Process Dimensions and Capabilities to Leverage Data-Driven Organizations Towards Industry 5.0. Appl. Syst. Innov. 2025, 8, 41. [Google Scholar] [CrossRef]
Figure 1. Correlation heatmap.
Figure 1. Correlation heatmap.
Systems 13 00773 g001
Figure 2. Descriptive distributions (density) of KD, PD, BDT, and DDT.
Figure 2. Descriptive distributions (density) of KD, PD, BDT, and DDT.
Systems 13 00773 g002
Table 1. Correlation matrix of variables.
Table 1. Correlation matrix of variables.
SizeTMTMOROEFLGSEFCDExitKDPDBDTDDT
Size10.419 **−0.401 **−0.0050.433 **0.407 **0.0060.054 **0.039 **−0.142 **0.075 **0.090 **0.050 **
TMT0.419 **1−0.219 **−0.0090.177 **0.327 **−0.082 **0.062 **−0.019 *−0.091 **−0.0160.107 **0.065 **
MO−0.401 **−0.219 **1−0.01−0.240 **−0.081 **0.011−0.012−0.028 **0.054 **−0.054 **0.094 **0.103 **
ROE−0.005−0.009−0.0110.026 **0.005−0.0070.0110.022 *−0.0030.007−0.019 *−0.007
FL0.433 **0.177 **−0.240 **0.026 **10.082 **−0.019 *0.031 **0.110 **−0.110 **0.016−0.016−0.082 **
GS0.407 **0.327 **−0.081 **0.0050.082 **1−0.057 **0.097 **−0.033 **−0.081 **−0.0050.234 **0.290 **
EF0.006−0.082 **0.011−0.007−0.019 *−0.057 **1−0.443 **−0.038 **0.485**0.325**−0.032**−0.013
CD0.054 **0.062 **−0.0120.0110.031 **0.097 **−0.443 **1−0.039 **−0.221 **0.039 **0.056 **0.074 **
Exit0.039 **−0.019 *−0.028 **0.022 *0.110 **−0.033 **−0.038 **−0.039 **1−0.105 **−0.086 **0.022 *−0.057 **
KD−0.142 **−0.091 **0.054 **−0.003−0.110 **−0.081 **0.485 **−0.221 **−0.105 **10.448 **−0.042 **−0.026 **
PD0.075 **−0.016−0.054 **0.0070.016−0.0050.325 **0.039 **−0.086 **0.448 **1−0.034 **−0.021 *
BDT0.090 **0.107 **0.094 **−0.019 *−0.0160.234 **−0.032 **0.056 **0.022 *−0.042 **−0.034 **10.545 **
DDT0.050 **0.065 **0.103 **−0.007−0.082 **0.290 **−0.0130.074 **−0.057 **−0.026 **−0.021 *0.545 **1
Mean23.10916.5000.0990.1090.4810.12671.89512.2880.15013.04517.2681.88015.580
SD1.3743.7610.1583.7770.2170.3117.6835.5960.35813.5998.3511.44133.041
Note: Firm size (Size); total number of top management team (TMT); management ownership (MO); return on equity (ROE); financial leverage (FL); government subsidies (GS); economic freedom (EF); cultural distance (CD); subsidiary exit (Exit); knowledge distance (KD); political distance (PD); breadth of digital transformation (BDT); depth of digital transformation (DDT); * p < 0.05, ** p < 0.01.
Table 2. Regression results.
Table 2. Regression results.
Exit
Model 1Model 2Model 3Model 4Model 5Model 6Model 7
Size−0.181−0.169−0.185−0.167−0.185−0.149−0.161
(0.029)(0.029)(0.029)(0.029)(0.029)(0.029)(0.029)
TMT0.0320.0350.0310.0340.0330.0310.028
(0.004)(0.004)(0.004)(0.004)(0.004)(0.004)(0.004)
MO−0.015−0.022−0.015−0.022−0.015−0.025−0.018
(0.103)(0.105)(0.102)(0.104)(0.102)(0.106)(0.103)
ROE0.0210.0190.0210.0200.0230.0190.021
(0.002)(0.002)(0.002)(0.002)(0.002)(0.002)(0.002)
FL0.108 ***0.111 ***0.107 ***0.110 ***0.105 **0.107 **0.101 **
(0.066)(0.068)(0.066)(0.067)(0.067)(0.069)(0.067)
GS−0.010−0.008−0.013−0.009−0.0130.001−0.003
(0.040)(0.039)(0.041)(0.040)(0.040)(0.038)(0.039)
EF0.1040.0200.0650.0100.0510.0290.067
(0.005)(0.003)(0.004)(0.004)(0.005)(0.003)(0.004)
CD0.0130.0320.0200.0350.0280.0410.032
(0.004)(0.003)(0.004)(0.003)(0.004)(0.004)(0.004)
KD −0.533 −0.610 −0.534
(0.011) (0.012) (0.011)
KD2 0.409 ** 0.449 *** 0.406 **
(<0.001) (<0.001) (<0.001)
PD 0.064 0.062 0.059
(0.003) (0.003) (0.003)
PD2 0.110 *** 0.098 *** 0.107 ***
(<0.001) (<0.001) (<0.001)
BDT −0.0320.073 **
(0.009)(0.008)
BDT_KD −0.021
(<0.001)
BDT_KD2 0.072 ***
(<0.001)
BDT_PD −0.022 ***
(<0.001)
BDT_PD2 −0.081 ***
(<0.001)
DDT −0.063 **−0.058 **
(<0.001)(<0.001)
DDT_KD 0.048 ***
(<0.001)
DDT_KD2 −0.051 ***
(<0.001)
DDT_PD 0.002
(<0.001)
DDT_PD2 −0.052 ***
(<0.001)
FE firmYesYesYesYesYesYesYes
FE host countryYesYesYesYesYesYesYes
FE yearYesYesYesYesYesYesYes
R20.2620.2710.2660.2720.2690.2730.268
R2_a0.1880.1980.1920.1990.1950.1990.194
F2.10931.64813.81235.35623.86826.42820.833
Observations12,19312,19312,19312,19312,19312,19312,193
Note: Standardized beta coefficients. Fixed effects: firm, host country, year. Multi-way clustered standard errors (firm and host country) in parentheses. ** p < 0.05, *** p < 0.01.
Table 3. Heterogeneity analysis.
Table 3. Heterogeneity analysis.
Exit
Developed EconomiesDeveloping Economies
Model 1Model 2Model 3Model 4Model 5Model 6Model 7Model 8
Size−0.197 *−0.211 *−0.185−0.194−0.272−0.273−0.219−0.214
(0.030)(0.030)(0.030)(0.029)(0.060)(0.058)(0.061)(0.059)
TMT−0.002−0.006−0.004−0.0110.1190.1120.1130.109
(0.004)(0.004)(0.004)(0.004)(0.010)(0.010)(0.010)(0.010)
MO−0.011−0.003−0.017−0.009−0.028−0.025−0.021−0.024
(0.122)(0.120)(0.123)(0.120)(0.189)(0.190)(0.209)(0.198)
ROE0.0200.0220.0200.0220.028 **0.032 **0.024 *0.024 *
(0.003)(0.003)(0.003)(0.003)(0.001)(0.001)(0.001)(0.001)
FL0.170 ***0.163 ***0.169 ***0.161 ***0.072 ***0.069 ***0.067 ***0.065 ***
(0.076)(0.078)(0.074)(0.077)(0.025)(0.025)(0.024)(0.024)
GS−0.009−0.014−0.001−0.007−0.003−0.0060.0140.014
(0.048)(0.049)(0.047)(0.047)(0.032)(0.033)(0.029)(0.030)
EF0.0330.1670.0560.1760.029−0.0070.0230.016
(0.006)(0.008)(0.006)(0.008)(0.004)(0.004)(0.006)(0.004)
CD0.0370.0200.0360.0200.0370.0210.0400.032
(0.006)(0.007)(0.006)(0.007)(0.004)(0.004)(0.003)(0.004)
KD−0.553 −0.507 4.189 1.805
(0.013) (0.012) (42.076) (6.239)
KD20.483 ** 0.446 ** 4.144 1.838
(<0.001) (<0.001) (1.878) (0.307)
PD 0.085 0.080 −0.090 −0.084
(0.003) (0.003) (0.004) (0.004)
PD2 0.077 ** 0.086 ** 0.020 0.026
(<0.001) (<0.001) (<0.001) (<0.001)
BDT−0.0400.051 −227.1500.090 *
(0.010)(0.009) (122.771)(0.012)
BDT_KD−0.051 ** −470.362
(<0.001) (21.806)
BDT_KD20.098 *** −243.181
(<0.001) (0.968)
BDT_PD −0.020 * −0.041 **
(<0.001) (<0.001)
BDT_PD2 −0.061 *** −0.081 **
(<0.001) (<0.001)
DDT −0.075 **−0.068 ** −72.132 **−0.084 **
(<0.001)(<0.001) (0.292)(<0.001)
DDT_KD 0.038 ** −150.898 **
(<0.001) (0.054)
DDT_KD2 −0.050 *** −78.883 **
(<0.001) (0.002)
DDT_PD −0.002 −0.002
(<0.001) (<0.001)
DDT_PD2 −0.039 ** −0.036 **
(<0.001) (<0.001)
FE firmYesYesYesYesYesYesYesYes
FE host countryYesYesYesYesYesYesYesYes
FE yearYesYesYesYesYesYesYesYes
R20.2830.2780.2830.2780.3550.3550.3540.354
R2_a0.2020.1970.2020.1970.2210.2220.2200.220
F107.73110.6243.260149.3218.3411.8224.1824.85
Observations94999499949994992694269426942694
Note: Standardized beta coefficients. Fixed effects: firm, host country, year. Multi-way clustered standard errors (firm and host country) in parentheses. * p < 0.1, ** p < 0.05, *** p < 0.01.
Table 4. Robustness test: replacing the econometric model (Cox).
Table 4. Robustness test: replacing the econometric model (Cox).
Exit
Model 1Model 2Model 3Model 4Model 5Model 6Model 7
Size0.098 ***0.114 ***0.097 ***0.110 ***0.093 **0.109 ***0.088 **
(0.033)(0.036)(0.033)(0.035)(0.033)(0.035)(0.033)
TMT−0.041−0.047−0.045−0.048−0.048−0.047−0.045
(0.013)(0.012)(0.013)(0.012)(0.013)(0.012)(0.013)
MO−0.024−0.029−0.028−0.036−0.035−0.015−0.019
(0.345)(0.322)(0.330)(0.315)(0.324)(0.315)(0.326)
ROE0.020 ***0.021 ***0.021 ***0.022 ***0.022 ***0.020 ***0.020 ***
(0.002)(0.002)(0.002)(0.002)(0.002)(0.002)(0.002)
FL0.078 ***0.075 ***0.077 ***0.074 ***0.082 ***0.074 ***0.078 ***
(0.073)(0.076)(0.074)(0.076)(0.074)(0.075)(0.073)
GS−0.118 **−0.137 **−0.126 **−0.141 **−0.128 **−0.089−0.078
(0.231)(0.243)(0.237)(0.249)(0.241)(0.230)(0.222)
EF−0.103 ***0.209 ***−0.0190.208 ***−0.0190.213 ***−0.018
(0.004)(0.005)(0.004)(0.005)(0.004)(0.005)(0.004)
CD−0.126 ***−0.216 ***−0.083 ***−0.216 ***−0.084 ***−0.209 ***−0.076 ***
(0.006)(0.005)(0.006)(0.005)(0.006)(0.005)(0.006)
KD −1.092 *** −1.101 *** −1.082 ***
(0.006) (0.006) (0.006)
KD2 0.870 *** 0.883 *** 0.851 ***
(<0.001) (<0.001) (<0.001)
PD −0.143 *** −0.167 *** −0.169 ***
(0.004) (0.004) (0.005)
PD2 0.113 *** 0.061 ** 0.065 **
(<0.001) (<0.001) (0.001)
BDT −0.0540.120 ***
(0.052)(0.033)
BDT_KD −0.040
(0.003)
BDT_KD2 0.120 *
(<0.001)
BDT_PD −0.050 **
(0.002)
BDT_PD2 −0.130 ***
(<0.001)
DDT −0.042−0.043
(0.002)(0.002)
DDT_KD 0.081 *
(<0.001)
DDT_KD2 −0.148 **
(<0.001)
DDT_PD −0.074
(<0.001)
DDT_PD2 −0.176 **
(<0.001)
FE yearYesYesYesYesYesYesYes
Observations12,19312,19312,19312,19312,19312,19312,193
Note: Standardized beta coefficients. Fixed effects: year. Multi-way clustered standard errors (firm and host country) in parentheses. * p < 0.1, ** p < 0.05, *** p < 0.01.
Table 5. Robustness test: replacing BDT and DDT.
Table 5. Robustness test: replacing BDT and DDT.
Exit
Model 1Model 2Model 3Model 4Model 5Model 6Model 7
Size−0.181−0.169−0.185−0.164−0.183−0.164−0.180
(0.029)(0.029)(0.029)(0.029)(0.029)(0.029)(0.029)
TMT0.0320.0350.0310.0350.0310.0340.031
(0.004)(0.004)(0.004)(0.004)(0.004)(0.004)(0.004)
MO−0.015−0.022−0.015−0.023−0.016−0.018−0.014
(0.103)(0.105)(0.102)(0.108)(0.104)(0.105)(0.103)
ROE0.0210.0190.0210.0200.0230.0190.022
(0.002)(0.002)(0.002)(0.002)(0.002)(0.002)(0.002)
FL0.108 ***0.111 ***0.107 ***0.110 ***0.105 **0.108 **0.103 **
(0.066)(0.068)(0.066)(0.067)(0.066)(0.070)(0.068)
GS−0.010−0.008−0.013−0.009−0.012−0.005−0.008
(0.040)(0.039)(0.041)(0.040)(0.041)(0.040)(0.042)
EF0.1040.0200.0650.0190.0670.0290.067
(0.005)(0.003)(0.004)(0.003)(0.004)(0.003)(0.004)
CD0.0130.0320.0200.0310.0170.0340.021
(0.004)(0.003)(0.004)(0.003)(0.004)(0.003)(0.004)
KD −0.533 −0.526 −0.531
(0.011) (0.011) (0.011)
KD2 0.409 ** 0.407 ** 0.409 **
(<0.001) (<0.001) (<0.001)
PD 0.064 0.058 0.060
(0.003) (0.003) (0.003)
PD2 0.110 *** 0.102 *** 0.098 ***
(<0.001) (<0.001) (<0.001)
BDT −0.0230.057 **
(0.007)(0.010)
BDT_KD 0.001
(<0.001)
BDT_KD2 0.048 ***
(<0.001)
BDT_PD −0.010
(<0.001)
BDT_PD2 −0.058 ***
(<0.001)
DDT −0.017−0.004
(0.011)(0.011)
DDT_KD 0.058 ***
(<0.001)
DDT_KD2 −0.048 **
(<0.001)
DDT_PD −0.002
(0.001)
DDT_PD2 −0.044 ***
(<0.001)
FE firmYesYesYesYesYesYesYes
FE host countryYesYesYesYesYesYesYes
FE yearYesYesYesYesYesYesYes
R20.2620.2710.2660.2720.2670.2720.267
R2_a0.1880.1980.1920.1990.1930.1990.193
F2.10931.64813.81234.51856.92725.96016.622
Observations12,19312,19312,19312,19312,19312,19312,193
Note: Standardized beta coefficients. Fixed effects: firm, host country, year. Multi-way clustered standard errors (firm and host country) in parentheses. ** p < 0.05, *** p < 0.01.
Table 6. Robustness test: excluding special industries.
Table 6. Robustness test: excluding special industries.
Exit
Model 1Model 2Model 3Model 4Model 5Model 6Model 7
Size−0.155−0.149−0.158−0.145−0.161−0.140−0.148
(0.041)(0.040)(0.041)(0.040)(0.041)(0.040)(0.040)
TMT0.0430.0450.0430.0440.0440.0410.041
(0.004)(0.005)(0.004)(0.005)(0.004)(0.005)(0.004)
MO−0.022−0.029−0.020−0.030−0.019−0.032−0.023
(0.129)(0.135)(0.129)(0.134)(0.127)(0.137)(0.129)
ROE0.055 **0.051 *0.054 **0.052 *0.054 **0.051 *0.054 **
(0.008)(0.007)(0.007)(0.007)(0.007)(0.007)(0.008)
FL0.168 ***0.175 ***0.166 ***0.172 ***0.165 ***0.175 ***0.164 ***
(0.062)(0.061)(0.062)(0.062)(0.062)(0.061)(0.062)
GS0.0180.0200.0160.0180.0160.0210.017
(0.041)(0.039)(0.041)(0.040)(0.041)(0.039)(0.041)
EF0.1000.0240.0600.0110.0450.0270.052
(0.005)(0.004)(0.005)(0.005)(0.005)(0.004)(0.005)
CD0.0260.0370.0370.0390.0450.0470.051
(0.005)(0.004)(0.005)(0.004)(0.005)(0.004)(0.005)
KD −0.625 −0.718 −0.597
(0.018) (0.018) (0.018)
KD2 0.433 * 0.496 ** 0.409 *
(<0.001) (<0.001) (<0.001)
PD 0.074 0.067 0.064
(0.003) (0.003) (0.003)
PD2 0.120 *** 0.092 *** 0.097 ***
(<0.001) (<0.001) (<0.001)
BDT −0.0570.044
(0.009)(0.009)
BDT_KD −0.038
(0.001)
BDT_KD2 0.089 ***
(<0.001)
BDT_PD −0.023 **
(<0.001)
BDT_PD2 −0.064 ***
(<0.001)
DDT −0.058−0.055
(0.001)(0.001)
DDT_KD 0.031
(<0.001)
DDT_KD2 −0.042 *
(<0.001)
DDT_PD −0.013
(<0.001)
DDT_PD2 −0.042 *
(<0.001)
FE firmYesYesYesYesYesYesYes
FE host countryYesYesYesYesYesYesYes
FE yearYesYesYesYesYesYesYes
R20.2600.2690.2650.2700.2660.2700.266
R2_a0.1830.1920.1870.1930.1890.1930.189
F6.54766.14713.74214.2115.52485.77216.184
Observations9175917591759175917591759175
Note: Standardized beta coefficients. Fixed effects: firm, host country, year. Multi-way clustered standard errors (firm and host country) in parentheses. * p < 0.1, ** p < 0.05, *** p < 0.01.
Table 7. PSM balance diagnostics.
Table 7. PSM balance diagnostics.
Balance Metrics (SMD)
Before MatchingAfter MatchingAfter Matching
Size−0.1080.0060.006
TMT0.0540.0010.001
MO0.0790.0010.001
ROE−0.061<0.001<0.001
FL−0.3100.0140.014
GS0.0910.0050.005
EF0.1050.0300.03
CD0.1080.0040.004
KD20.0360.049
PD20.208 0.018
Table 8. Robustness check: Propensity Score Matching (PSM).
Table 8. Robustness check: Propensity Score Matching (PSM).
Exit
Model 1Model 2
Size−0.142−0.149
(0.031)(0.030)
TMT0.0190.015
(0.004)(0.004)
MO−0.014−0.007
(0.105)(0.100)
ROE0.0190.021
(0.002)(0.002)
FL0.114 **0.112 **
(0.057)(0.056)
GS−0.014−0.020
(0.034)(0.034)
EF0.0180.062
(0.003)(0.004)
CD0.0390.022
(0.004)(0.004)
KD−0.567
(0.011)
KD20.413 ***
(<0.001)
PD 0.055
(0.002)
PD2 0.111 ***
(<0.001)
FE firmYesYes
FE host countryYesYes
FE yearYesYes
R20.2760.271
R2_a0.2030.198
F25.94013.003
Observations12,19312,193
Note: Standardized beta coefficients. Fixed effects: firm, host country, year. Multi-way clustered standard errors (firm and host country) in parentheses. ** p < 0.05, *** p < 0.01.
Table 9. Endogeneity test: nonlinear control function method (2SRI).
Table 9. Endogeneity test: nonlinear control function method (2SRI).
Exit
Model 1Model 2
Size−0.171−0.184
(0.029)(0.029)
TMT0.0360.034
(0.004)(0.004)
MO−0.018−0.016
(0.105)(0.100)
ROE0.0190.021
(0.002)(0.002)
FL0.111 ***0.108 ***
(0.068)(0.065)
GS−0.009−0.014
(0.039)(0.041)
EF0.0210.062
(0.003)(0.005)
CD0.0370.010
(0.003)(0.004)
KD−0.149
(0.011)
KD21.159
(0.002)
Resid_KD−0.071
(0.015)
Resid_KD2−0.327
(0.002)
PD −0.541
(0.021)
PD2 2.815
(0.022)
Resid_PD 0.141
(0.021)
Resid_PD2 −1.290
(0.022)
FE firmYesYes
FE host countryYesYes
FE yearYesYes
R20.2720.266
R2_a0.1980.192
F26.41011.396
Observations12,19312,193
Note: Standardized beta coefficients. Fixed effects: firm, host country, year. Multi-way clustered standard errors (firm and host country) in parentheses. *** p < 0.01.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Zhou, Z.; Wang, L. Dual-Dimensional Digital Transformation Systematically Reshapes the U-Curve of Knowledge and Political Distance on Subsidiary Exit. Systems 2025, 13, 773. https://doi.org/10.3390/systems13090773

AMA Style

Zhou Z, Wang L. Dual-Dimensional Digital Transformation Systematically Reshapes the U-Curve of Knowledge and Political Distance on Subsidiary Exit. Systems. 2025; 13(9):773. https://doi.org/10.3390/systems13090773

Chicago/Turabian Style

Zhou, Zhengyuan, and Lei Wang. 2025. "Dual-Dimensional Digital Transformation Systematically Reshapes the U-Curve of Knowledge and Political Distance on Subsidiary Exit" Systems 13, no. 9: 773. https://doi.org/10.3390/systems13090773

APA Style

Zhou, Z., & Wang, L. (2025). Dual-Dimensional Digital Transformation Systematically Reshapes the U-Curve of Knowledge and Political Distance on Subsidiary Exit. Systems, 13(9), 773. https://doi.org/10.3390/systems13090773

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop