ESG Performance and the Phase-Dependent Resilience of Outward Foreign Direct Investment: Evidence from Chinese Multinationals

Abstract

Chinese multinational enterprises, as the most active emerging-market investors, face mounting challenges in sustaining outward foreign direct investment (OFDI) under increasingly volatile global environments, yet how ESG performance shapes firms’ capacity to withstand and recover from external shocks remains poorly understood. This study investigates whether and how ESG performance enhances the OFDI resilience of Chinese multinational enterprises across the resistance phase and the recovery phase. We hypothesize that ESG performance enhances OFDI resilience through phase-specific mechanisms: in the resistance phase, ESG functions as a static resource buffer grounded in the resource-based view, while in the recovery phase, it operates as a dynamic reconfiguration mechanism consistent with the dynamic capabilities view. Using a panel dataset of 19,691 firm-year observations from Chinese A-share listed firms spanning 2008 to 2024, we employ a fixed-effects panel model to test these hypotheses. The results show that ESG performance significantly enhances OFDI resilience in both phases, and this conclusion holds after robustness and endogeneity tests. Mechanism analysis reveals that green innovation mediates the effect in both the resistance and recovery phases, while supply chain resilience and investment efficiency serve as additional mediating channels exclusively in the resistance phase. By introducing a phase-dependent perspective and highlighting ESG’s distinct roles across shock stages, this study provides practical guidance for emerging-market multinational enterprises on how to leverage ESG performance to build sustainable OFDI resilience in volatile global environments.

1. Introduction

With the resurgence of cross-border investment protectionism, the global investment environment has become increasingly uncertain. Host countries have systematically tightened outward foreign direct investment (OFDI) entry requirements through enhanced investment screening mechanisms, an expanded scope of national security reviews, and sector-specific restrictions in selected technology-intensive industries, thereby reducing the openness and predictability of the cross-border capital flows [1]. Under a growing emphasis on security and resilience, global supply chains are accelerating toward regionalization and localization, reshaping both the location choices and organizational structures of OFDI [2]. This shift not only raises investment costs but also imposes significant constraints on the long-term continuity and stability of OFDI. How to maintain the continuity and stability of Chinese firms’ OFDI has become a critical concern for both entrepreneurs and policymakers.

A substantial body of research has focused on improving the continuity and stability of OFDI from the perspectives of institutional quality [3], political and geopolitical stability [4], macroeconomic and financial stability [5], and reinvestment [6]. Similarly, researchers have investigated the effects of continuity and stability of OFDI on economic growth and stability [7], knowledge diffusion [8], and low-carbon transition [9]. More recently, an emerging stream of research has begun to investigate the resilience of OFDI under external shocks. This line of research examines the role of artificial intelligence in shaping OFDI resilience from the perspective of enterprise heterogeneity [10], as well as aggregate FDI resilience in the post-COVID period [11]. These studies largely overlook the role of firm-level governance factors in shaping OFDI resilience and tend to conceptualize resilience as a unidimensional construct without distinguishing between the resistance and recovery phases.

The concept of resilience, which originally stems from the fields of ecology and engineering [12], has been widely introduced in economics to describe the capacity of economic systems to resist and recover from external shocks [13]. In this context, we define OFDI resilience as the capacity of remaining stable or recovering quickly in response to shocks, ensuring sustained capital inflows, reinvestment of earnings, and ongoing productive activities in the host country. OFDI resilience not only reflects firms’ ability to withstand external shocks, but also emphasizes the dynamic adaptive capacity to continuously adjust, recover, and reconfigure OFDI activities in complex environments.

The OFDI resilience depends not only on the external institutional environment but also closely relates to firms’ endogenous governance capabilities in the uncertain international environment. Compared with traditional firm characteristics such as size, age, financial structure, ownership structure, board characteristics, and international experience, Environmental, Social and Governance (ESG) performance more directly reflects firms’ ability to adapt to institutional complexity [14], stakeholder pressures [15], and environmental uncertainty [16]. The ESG framework, by integrating environmental governance, social responsibility performance, and corporate governance quality into strategic decision-making, provides potential support for enhancing FDI resilience.

Despite growing recognition of ESG as a source of long-term competitive advantage, including through the integration of digital technologies that enhance ESG performance via correlation and spillover effects [17], its role in enhancing OFDI resilience remains strikingly underexplored. While prior studies have linked ESG to general firm performance [18], internationalization propensity [19], or green OFDI [20], and some recent studies have begun to explore ESG’s contribution to corporate or organizational resilience in domestic or crisis contexts [21], most of these studies treat resilience as a unidimensional construct and do not systematically distinguish between the resistance phase and the recovery phase. Little research has systematically examined whether and how ESG performance contributes to OFDI resilience in the face of external shocks. This constitutes an important theoretical and practical gap, because external shocks unfold over time. Firms need immediate buffering capacity in the short term, as well as legitimacy restoration and resource reconfiguration capacity in the longer term. ESG performance may contribute to these capacities in different ways across stages. The present study addresses this gap by explicitly distinguishing the resistance phase and the recovery phase, thereby providing a more nuanced understanding of how ESG performance enhances OFDI resilience under uncertainty.

Building on the above gap, this study examines whether ESG performance enhances Chinese firms’ OFDI resilience and how this effect varies across the resistance and recovery phases. Drawing on the resource-based view and institutional theory, we argue that superior ESG performance equips firms with reputational capital, stakeholder trust, and adaptive governance capabilities, which serve as short-term buffering mechanisms during the resistance phase and as legitimacy repair and resource reconfiguration mechanisms during the recovery phase. We further identify three mediating channels, namely green innovation, supply chain resilience, and investment efficiency, as well as three boundary conditions, including industry environmental sensitivity, degree of digital transformation, and host-country income level. In doing so, this study offers a more nuanced understanding of how ESG practices contribute to OFDI resilience under external shocks.

Using a large panel dataset of Chinese A-share listed firms’ OFDI activities from 2008 to 2024 and employing a fixed-effects panel model, this study shows that ESG performance positively affects OFDI resilience in both the resistance and recovery phases. Phase-dependent boundary conditions are observed, and mediation analyses confirm multiple transmission channels through which ESG enhances resilience. These results are robust to a range of checks, including sample refinement, winsorization, and alternative estimation methods.

This study makes several contributions. First, it introduces and systematically examines the concept of OFDI resilience, a critical yet underexplored dimension in OFDI research, and shows that ESG performance significantly contributes to OFDI resilience under external shocks. By explicitly distinguishing between the resistance and recovery phases, this study reveals that ESG functions as a short-term buffering mechanism during acute disruptions and as a long-term legitimacy repair and resource reconfiguration mechanism during recovery, thereby extending the traditional ESG–performance nexus into the dynamic domain of international investment resilience.

Second, it provides evidence from an emerging-market context by showing how Chinese firms leverage ESG practices to overcome liabilities of foreignness and origin in volatile global environments. The phase-dependent boundary conditions, including stronger effects in low-environmental-sensitivity industries and high-digital-transformation firms during the resistance phase, and in high-environmental-sensitivity industries during the recovery phase, offer actionable insights for emerging-market multinationals navigating geopolitical and regulatory uncertainties.

Third, by integrating three theoretically grounded mediating channels and three key boundary conditions, this study develops a comprehensive and nuanced theoretical framework that explains not only whether, but also when and how, ESG performance translates into OFDI resilience. This multi-path, phase-dependent approach advances our understanding of the micro-foundations of resilience in international business and provides a solid basis for future research on strategic resources in uncertain global contexts.

The remainder of the paper is structured as follows: Section 2 develops the theoretical background and hypotheses, Section 3 describes the data and methodology, Section 4 presents the empirical results, and Section 5 concludes with a discussion of the findings, implications, limitations, and future research directions.

2. Theoretical Background and Hypotheses

2.1. ESG Performance and OFDI Resilience

The Eclectic Paradigm (OLI framework) provides a foundational lens for understanding why firms invest abroad. It suggests that firms tend to achieve cross-border expansion by leveraging ownership (O), location (L), and internalization (I) advantage [22]. While the OLI paradigm effectively captures the initial motivation for cross-border investment, it offers limited insight into how firms sustain and renew these advantages amid volatile host-country institutional environments and external shocks. To address this limitation, we draw on the resource-based view (RBV), which argues that sustained competitive advantage derives from valuable, rare, imperfectly imitable, and non-substitutable (VRIN) firm-specific resources and capabilities [23]. In the OFDI context, superior ESG performance constitutes a modern, intangible ownership advantage under the RBV, embedded in firm-specific governance routines, stakeholder relationships, and socially embedded relational capital, and thus difficult to replicate [24]. Yet, the RBV’s static focus on resource stocks constrains its explanatory power when host-country environments are characterized by institutional volatility, regulatory shifts, and sudden disruptions—conditions under which pre-existing advantages may erode without active renewal.

We therefore extend the framework with the dynamic capabilities view (DCV), which posits that competitive advantage in turbulent settings depends on managerial processes to sense emerging opportunities, seize them, and reconfigure internal and external resources through continuous learning and innovation [25]. By hierarchically integrating OLI’s ownership advantage logic, the resource-based view’s focus on firm-specific resource endowments, and the dynamic capabilities view’s emphasis on ongoing resource renewal, this study develops a unified analytical lens to explain how ESG performance enhances OFDI resilience. Specifically, ESG performance originates as an ownership advantage, serves as a static buffer against immediate shocks, and subsequently evolves into a dynamic reconfiguration capability that supports long-term adaptation.

To delineate the boundaries and differential explanatory power of these theories, we incorporate a two-stage resilience perspective drawn from Martin and Sunley (2015) [26], which distinguishes: (1) the resistance phase, where firms must absorb immediate shocks and preserve core operations under sudden disruptions, and (2) the recovery phase, where firms adapt to altered environments, release locked resources, and pursue new growth paths through reconfiguration and innovation.

In the resistance phase, ESG performance primarily operates as a static buffer under the RBV. The term static denotes that firms deploy pre-accumulated ESG-related resources in response to shocks, rather than actively reconfiguring them in real time. During the initial stage of disruption, severe time constraints and heightened uncertainty often limit firms’ ability to engage in dynamic capability deployment, rendering immediate resource reconfiguration infeasible. High ESG performance signals legitimacy and responsibility to host-country stakeholders, thereby reducing reputational frictions and facilitating the accumulation of trust-based relationships with local governments, communities, and partners [27]. Over time, these relationships evolve into relational capital that is embedded within the firm’s existing resource base [28]. When disruptions occur, such capital can be mobilized to absorb shocks by alleviating resource constraints, reducing transaction and coordination costs, and stabilizing stakeholder support, thereby mitigating operational disruptions and sustaining production. This buffering effect relies on ESG-related resources that have been accumulated over time and can be deployed as a pre-existing stock, rather than actively reconfigured in response to shocks. Consistent with the RBV, such resources may exhibit VRIN characteristics and thus serve as a primary source of resilience when rapid adaptation is constrained.

In contrast, the recovery phase requires active adaptation and path creation, during which firms with superior ESG performance develop dynamic reconfiguration capability under the DCV. ESG engagement facilitates the development of dynamic capabilities, particularly absorptive and adaptive capabilities, which are critical for recovery and long-term adjustment [29]. On the one hand, absorptive capability enables firms to identify, assimilate, and apply external knowledge regarding regulatory changes and shifting stakeholder expectations, thereby facilitating the recognition of sustainability-oriented opportunities [30]. On the other hand, adaptive capability allows firms to adjust strategic responses and mobilize resources in changing environments, supporting the effective pursuit of these opportunities through resource realignment and organizational adjustment [31]. Through this process, firms move beyond mere recovery by abandoning obsolete routines, recombining existing resources, and developing new sources of competitive advantage. Such transformation reflects a form of path creation, whereby firms not only restore pre-shock performance but also establish new development trajectories that sustain long-term adaptability in international markets. Accordingly, resilience in the recovery phase depends less on passive buffering and more on continuous renewal and proactive strategic repositioning. This study conceptualizes resilience as a stage-dependent process, in which ESG performance functions as a static buffer in the resistance phase and as a dynamic reconfiguration capability in the recovery phase. On this basis, we propose the following hypotheses:

H1a: 

ESG performance exerts a significant positive effect on the resilience of outward foreign direct investment during the resistance phase.

H1b: 

ESG performance exerts a significant positive effect on the resilience of outward foreign direct investment during the recovery phase.

2.2. The Mediating Effect of Green Innovation

Beyond direct competitive advantages, ESG performance may enhance OFDI resilience through green innovation. In the resistance phase, green innovation mediates the relationship between ESG performance and OFDI resilience via a resource-stock mechanism rooted in the RBV. Strong ESG performance alleviates financing constraints and reduces agency costs, channeling organizational resources toward long-term green R&D investment rather than short-term profit extraction [32]. This sustained redirection of resources generates, over time, a pre-accumulated stock of green technological assets—clean production processes, low-carbon product portfolios, and environmental certifications. Such assets are precisely what the RBV identifies as VRIN resources: scarce, difficult to imitate, and deeply embedded in firms’ operational routines in ways that rivals cannot easily replicate. It is this inimitability that determines their resilience value. When sudden regulatory shocks or geopolitical disruptions strike host-country operations, inimitable green assets allow firms to maintain regulatory compliance without costly emergency adjustments, preserve reputational credibility under heightened stakeholder scrutiny, and reduce dependence on conventional energy inputs—collectively enabling firms to absorb external shocks and sustain overseas operational continuity [33]. Sustaining continuity, rather than achieving growth, is the defining challenge of the resistance phase, and pre-accumulated green innovation constitutes the firm-specific buffer through which ESG performance meets that challenge.

In the recovery phase, however, this buffering logic reaches its boundary. As environmental uncertainty stabilizes and firms shift from disruption containment to strategic renewal, the static deployment of pre-existing assets is no longer sufficient—what is required is active reconfiguration, marking a theoretical transition from the RBV to the DCV. Reconfiguration, in turn, demands new knowledge inputs rather than the redeployment of existing stocks. ESG performance provides precisely these inputs by deepening firms’ embeddedness in external knowledge networks—spanning regulators, research institutions, and sustainability-oriented partners—that continuously feed new environmental knowledge into firms’ innovation processes [34]. These innovation processes, energized by ESG-driven organizational learning, generate dynamically renewed green innovation outputs aligned with evolving host-country institutional requirements. Aligned outputs matter because post-shock host-country environments are rarely identical to pre-shock ones: new regulatory standards, shifting stakeholder expectations, and emerging green market opportunities collectively redefine what constitutes competitive advantage [35]. Green innovation, continuously renewed through ESG engagement, provides the technological foundation for sensing these emerging opportunities, seizing them through new low-carbon products, and reconfiguring international investment portfolios toward new growth trajectories—moving firms beyond mere restoration of pre-shock performance toward genuinely renewed competitive positioning in international markets, consistent with the DCV’s emphasis on sensing, seizing, and transforming as the primary source of resilience in turbulent environments. On this basis, we propose the following hypotheses:

H2a: 

ESG performance enhances OFDI resilience through the mediating role of green innovation during the resistance phase.

H2b: 

ESG performance enhances OFDI resilience through the mediating role of green innovation during the recovery phase.

2.3. The Mediating Effect of Supply Chain Resilience

Strong ESG performance signals reliability and responsibility to external stakeholders, thereby making the firm more appealing to global partners and supporting deeper supply chain collaboration [36]. From the RBV perspective, the resulting high-quality supplier network constitutes a firm-specific resource that is valuable, rare, and difficult for rivals to replicate, embedding relational capital and governance routines that strengthen the firm’s strategic asset base. Although stringent ESG standards may initially screen out non-compliant suppliers, they simultaneously help build a value-aligned supplier network marked by deeper trust and smoother coordination [37]. This alignment enables firms to proactively develop diversified and redundant supply relationships, thereby reducing reliance on any single supplier and spreading operational risks more effectively [38].

In the resistance phase, these structural and relational capabilities allow firms to absorb sudden disruptions by quickly reconfiguring supply relationships, adjusting production schedules, and maintaining continuity of operations, thereby significantly mediating the effect of ESG performance on OFDI resilience. The governance (G) dimension of ESG plays a pivotal role in building supply chain resilience. Strong governance helps reduce information asymmetry by encouraging transparency and open sharing of information throughout the supply chain [39]. The governance practice also fosters the gradual accumulation of relational capital [40]. When external disruptions strike, this relational capital delivers concrete adaptive benefits. Suppliers are more willing to offer flexible payment terms, prioritize deliveries, and jointly adjust production schedules. Such collaborative responses enable firms to maintain operational continuity and production stability even in volatile host country environments.

In the recovery phase, the mediating role of supply chain resilience continues to persist, but its function evolves from shock absorption to enabling dynamic adaptation under the DCV. As environmental uncertainty begins to stabilize and firms shift from disruption management to strategic adjustment, supply chain resilience provides a critical foundation for the deployment of dynamic capabilities. Specifically, resilient supply chain structures ensure stable access to key inputs, timely information flows, and close coordination with partners, which are essential for firms to sense emerging opportunities and respond to changing institutional environments [41]. These conditions allow firms to reconfigure sourcing structures, adjust production networks, and redeploy resources across markets in response to post-shock changes. In this sense, supply chain resilience not only sustains operational continuity but also facilitates the resource reconfiguration and strategic adjustment required in the recovery phase.

By integrating structural flexibility with high-quality relational stability, ESG performance builds a supply chain system capable of absorbing shocks, adapting to institutional uncertainty, and recovering from disruptions. On this basis, we propose the following hypotheses:

H3a: 

ESG performance enhances OFDI resilience through the mediating role of supply chain resilience during the resistance phase.

H3b: 

ESG performance enhances OFDI resilience through the mediating role of supply chain resilience during the recovery phase.

2.4. The Mediating Effect of Investment Efficiency

ESG performance can enhance OFDI resilience by improving investment efficiency. Stronger ESG performance can improve internal governance by enhancing transparency, increasing stakeholder scrutiny, and strengthening board oversight [42]. In the absence of effective monitoring, managerial opportunism often leads to overinvestment and underinvestment, distorting capital allocation [43]. By aligning managerial incentives more closely with long-term shareholder interests, strong ESG performance fosters greater discipline and precision in investment decisions, thereby reducing wasteful expenditures and ensuring resources are directed toward high-value opportunities [44].

Beyond its internal governance benefits, ESG performance also functions as an external credibility signal in capital markets [45]. High ESG ratings strengthen firms’ reputational capital, attract long-term institutional investors, and expand access to green financing instruments at lower costs [46]. These effects ease financing constraints and enable more efficient deployment of capital across domestic and international projects.

In the resistance phase, improved investment efficiency operates as a static buffering mechanism under the RBV. Firms with disciplined capital allocation are less prone to inefficient or excessive investments, enabling them to preserve liquidity, maintain financial stability, and sustain the continuity of overseas operations during sudden disruptions such as regulatory changes or geopolitical tensions [47]. This efficient deployment of financial resources reduces exposure to investment risks and allows firms to absorb external shocks without immediate operational collapse, thereby strengthening OFDI resilience. In the recovery phase, investment efficiency supports resilience through dynamic resource reconfiguration under the DCV. As firms transition from shock absorption to adaptation and path creation, efficient capital allocation enables them to sense emerging opportunities in host countries, seize these prospects through targeted reallocation, and reconfigure international investment portfolios toward new growth trajectories (e.g., shifting to sustainable projects or diversifying markets). In this sense, investment efficiency not only sustains financial stability but also facilitates the strategic renewal required for recovery [48], thereby supporting firms’ ability to restore and enhance their OFDI activities in the long run. On this basis, we propose the following hypotheses:

H4a: 

ESG performance enhances OFDI resilience through the mediating role of investment efficiency during the resistance phase.

H4b: 

ESG performance enhances OFDI resilience through the mediating role of investment efficiency during the recovery phase.

3. Research Design

3.1. Sample and Data

We selected Chinese A-share listed firms in the Shanghai and Shenzhen stock exchanges from 2008 to 2024 as our initial sample. As the world’s largest emerging economy and one of the most active outward foreign direct investors, China provides an ideal setting for studying OFDI resilience. First, Chinese firms operate in a highly dynamic institutional environment characterized by frequent policy shifts, evolving regulatory frameworks, and significant geopolitical uncertainty [49]. These conditions make OFDI resilience particularly salient, as firms must simultaneously navigate home-country institutional pressures and host-country risks. Second, China has experienced rapid growth in outward foreign direct investment since the global financial crisis, accompanied by increasing exposure to external shocks such as trade frictions, supply chain disruptions, and regulatory tightening in host countries. This context allows us to observe how firms build and maintain resilience in real-world volatile conditions. Third, systematic disclosure of detailed firm-level OFDI and ESG data began around 2008 in China, ensuring both temporal coverage and data completeness for our analysis [50]. The sample period ends in 2024, the latest year with data available at the time of this study.

Firm-level data on OFDI activities, ESG performance, control variables, and mediating variables were primarily obtained from the CSMAR (China Stock Market & Accounting Research) and the Chinese Research Data Services (CNRDS) databases. Country-level control variables were sourced from the World Bank database. To ensure data quality and minimize potential estimation bias, the initial sample was subject to the following screening procedures: (1) firms designated as ST, PT, or *ST were excluded, as their abnormal financial conditions may distort empirical results; (2) firm-year observations with missing values in ESG scores, OFDI measures, or main control variables were dropped.

3.2. Measurement

OFDI resilience: The measurement of OFDI resilience in this study is grounded in the performance-deviation approach widely used in the economic resilience literature [26]. This approach conceptualizes resilience as the ability to absorb shocks and subsequently recover by comparing actual outcomes with a counterfactual trajectory that represents expected performance in the absence of shocks. Rather than relying solely on observed contractions and expansions, this framework evaluates resilience by assessing the extent to which actual investment dynamics deviate from their expected trend.

Prior studies construct counterfactual benchmarks by projecting pre-shock growth trends into shock and recovery periods, thereby approximating what would have occurred in the absence of external disturbances. The difference between actual and expected outcomes captures the magnitude of shock impact (resistance) and the extent of subsequent adjustment (recovery). This enables a unified and dynamic assessment of both resistance and recovery, while providing a theoretically grounded benchmark for evaluating firms’ responses to external shocks. Furthermore, when the focus is on how different firms respond to a common external shock, the expected trajectory can be benchmarked against aggregate dynamics. Under the assumption that all else equal, an individual firm would follow the overall trend in the absence of idiosyncratic frictions, the aggregate growth rate serves as a natural counterfactual reference. In this sense, the expected change in an individual firm can be approximated by applying the overall growth rate of the system to its pre-shock level.

Building on this logic, we extend the performance-deviation approach to the OFDI context by measuring resilience as the relative deviation of actual OFDI volumes from their trend-based counterfactual trajectories. We first construct each firm’s expected OFDI growth trajectory based on its historical pre-shock trends. The relative deviation is then computed as the difference between actual OFDI volumes and the counterfactual trajectory, scaled by the counterfactual value. A positive value indicates that actual growth exceeds expectations. A value of zero indicates that actual performance exactly matches the expected trajectory. A negative value indicates underperformance relative to expectations. The approach is essentially consistent with existing measures in the literature [10], differing only in that we explicitly separate resistance and recovery periods rather than treating resilience as a single undifferentiated measure.

From a theoretical perspective, this classification follows the economic resilience framework, which conceptualizes resilience as a dynamic process consisting of a resistance phase and a subsequent recovery phase. Importantly, recovery does not necessarily imply a full return to the pre-shock equilibrium, but rather a period of post-shock adjustment during which firms gradually adapt to new conditions. Moreover, the impact of major shocks is typically strongest at the onset and tends to attenuate over time as expectations partially stabilize and firms adjust their behavior. In this context, we designate 2008–2009 and 2018–2020 as resistance periods and 2010–2017 and 2021–2024 as recovery periods. The 2008–2009 period captures the immediate impact of the global financial crisis. The 2018–2020 period reflects the escalation of US–China trade tensions, which began in March 2018 following the U.S. Section 301 tariff announcement and continued until the signing of the Phase One economic and trade agreement in January 2020, combined with the onset of the COVID-19 pandemic. The specific calculation of OFDI resilience is as follows:

$${OFDIRes}_{\mathit{ijt}}={\displaystyle \frac{{\left({∆OFDI}_{\mathit{ijt}}^{resistance}\right)}^{observed}-{\left({∆OFDI}_{\mathit{ijt}}^{resistance}\right)}^{expected}}{\left|{\left({∆OFDI}_{\mathit{ijt}}^{resistance}\right)}^{expected}\right|}}$$

(1)

$${OFDIRec}_{\mathit{ijt}}={\displaystyle \frac{{\left({∆OFDI}_{\mathit{ijt}}^{recovery}\right)}^{observed}-{\left({∆OFDI}_{\mathit{ijt}}^{recovery}\right)}^{expected}}{\left|{\left({∆OFDI}_{\mathit{ijt}}^{recovery}\right)}^{expected}\right|}}$$

(2)

$${\left({∆OFDI}_{\mathit{ijt}}\right)}^{expected}=g_b\times {OFDI}_{{\mathit{ijt}}_0}$$

(3)

where $g_b$ represents the aggregate growth rate of OFDI across all firms in the sample during the resistance or recovery period and serves as a benchmark for OFDI dynamics, ${OFDI}_{{\mathit{ijt}}_0}$ denotes the level of OFDI undertaken by firm i in host country j at the base period $t_0$.

ESG performance: Given the institutional context of this study, we adopt the Chinese Research Data Services (CNRDS) ESG rating to measure corporate ESG performance [51]. The CNRDS rating system is characterized by relatively frequent updates, broad firm coverage, and high data reliability, making it particularly suitable for the Chinese market. Under this system, the ESG performance of listed firms is scored from 0 to 100, with higher values indicating better ESG performance.

Green innovation: Firm-level green innovation is typically assessed from input or output perspectives. However, as green innovation inputs are difficult to disentangle from general R&D expenditures, we measure green innovation from the output perspective using the number of green patent applications and the number of green patents granted [52].

Supply chain resilience: We capture supply chain resilience from the perspective of supply chain resistance using two proxy measures. Supply chain concentration is calculated as the simple average of (i) the proportion of total purchases accounted for by the firm’s top five suppliers and (ii) the proportion of total sales accounted for by the firm’s top five customers, thereby reflecting concentration in upstream and downstream relationships [53]. Higher values indicate greater dependence on a small number of suppliers and customers. Supply chain resistance is measured as the natural logarithm of the ratio of the sum of accounts receivable and advance payments to operating revenue [54]. Lower values suggest a stronger capacity of firms to withstand supply chain disruptions.

Investment efficiency: Consistent with the literature, investment inefficiency is operationalized using residual-based measures from investment expectation models. We adopt two established approaches. The first measure (InveEffic1) follows the model developed by Chen et al. (2011) [55]. In this model, firm investment in year t is regressed on lagged sales growth, an interaction between sales growth and an indicator for negative growth, and a set of firm-level control variables. The interaction term allows investment sensitivity to differ between expansion and contraction periods. Investment inefficiency is measured as the absolute value of the regression residual. The second measure (InveEffic2) follows Biddle et al. (2009) [56], in which investment is regressed on lagged sales growth and the same set of control variables, without distinguishing between positive and negative growth.

Control variables: We include a set of control variables at both the firm and country levels. Firm-level controls comprise firm size (Size), measured as the natural logarithm of the number of total assets plus one [57]; return on equity (ROE), defined as net profit divided by average shareholders’ equity [58]; ownership concentration (Top1), measured by the share proportion held by the largest tradable shareholder [59]; financing constraints (SA), proxied by the SA index [60]; and research and development intensity (Invent), measured as R&D expenditure relative to operating revenue [61].

In addition, we control for country-level characteristics that may influence OFDI resilience. These include economic size (GDP), measured as the natural logarithm of per capita GDP plus one [62]; the labor force participation rate (LP), defined as the ratio of the labor force to the working-age population [63]; institutional quality (WGI), captured by the Worldwide Governance Indicators [64]; and trade openness (Trade), measured as the share of imports of goods and services in GDP [65].

Table 1. Variables.

Variable	Symbol	Definition
OFDI resistance capacity	OFDIRes	As mentioned above
OFDI recovery capacity	OFDIRec	As mentioned above
ESG performance	ESG	CNRDS ESG Rating Index
Firm size	Size	Ln (total assets + 1)
Return on equity	ROE	Net income/average shareholders’ equity
Ownership concentration	Top1	Share proportion held by the largest tradable shareholder
Financial constraints	SA	SA index
R&D intensity	Invent	R&D expenditure as a share of operating revenue
Economic scale	GDP	Ln (GDP per capita + 1)
Labor force participation rate	LP	Labor force/working-age population
Institutional quality	WGI	Government effectiveness score
Trade openness	Trade	The share of imports of goods and services in GDP
Green innovation	Patent1	Number of green patent applications
	Patent2	Number of green patents granted
Supply chain resilience	SupResis1	Mean of the top-five supplier purchase share and top-five customer sales share
	SupResis2	Ln [(accounts receivable + advance payments)/operating revenue]
Investment efficiency	InveEffic1	As mentioned above
	InveEffic2	As mentioned above

reports detailed definitions of all variables.

3.3. Summary Statistics

Table 2. Descriptive statistics.

Variable	Obs	Mean	SD	Min	Max
OFDIRes	6849	0.0981	0.2864	−0.7541	0.8963.
OFDIRec	12,872	−0.0768	0.1221	−1.3062	1.1738
ESG	19,691	23.0749	9.0531	0.1637	75.6302
Size	19,691	22.9269	1.4894	19.2837	28.7005
ROE	19,691	0.0656	0.2974	−22.1201	7.3769
Top1	19,691	24.2100	17.6475	0.0029	89.0876
SA	19,691	−3.8057	0.4101	−5.8306	−0.5272
Invent	19,691	5.7673	6.6046	0.0000	77.3011
GDP	19,691	12.2840	1.8382	5.9805	19.0265
LP	19,691	65.0452	5.8374	38.7442	91.3928
WGI	19,691	1.0485	0.9282	−1.7037	2.4014
Trade	19,691	91.6163	76.0281	1.1342	221.0125
Patent1	19,691	7.2032	45.4582	0.0000	1595.0000
Patent2	19,691	3.6752	19.5583	0.0000	574.0000
SupResis1	19,691	0.1992	0.1343	0.0816	0.3921
SupResis2	19,691	−1.4378	0.8644	−9.7421	6.4232
InveEffic1	19,691	0.0341	0.0251	0.0000	0.1931
InveEffic2	19,691	0.0351	0.0273	0.0000	0.2943

reports the descriptive statistics for the variables in this study. The mean value of OFDI resilience during the resistance phase (OFDIRes) is 0.0981 (SD = 0.2864), whereas the mean value during the recovery phase (OFDIRec) is −0.0768 (SD = 0.1221). The relatively larger standard deviation in the resistance phase suggests that firms exhibit greater heterogeneity in their ability to withstand external shocks, while the lower dispersion in the recovery phase indicates more convergent post-shock adjustment paths. The mean ESG score is 23.0749 (SD = 9.0531), indicating considerable cross-sectional variation in ESG performance among Chinese listed firms. This variation provides a suitable empirical basis for examining the heterogeneous effects of ESG on OFDI resilience. For firm-level characteristics, Size shows moderate dispersion (SD = 1.4894), suggesting relatively stable firm scale across the sample, while return on equity exhibits substantial variability (SD = 0.2974, with extreme values ranging from −22.1201 to 7.3769), reflecting substantial differences in profitability and potential outliers. Ownership concentration also displays a wide range, indicating heterogeneous corporate governance structures. Financial constraints have a mean of −3.8057 with relatively low dispersion (SD = 0.4101), suggesting that most firms in the sample face a comparable level of financing constraints, although the presence of variation still allows for identifying its moderating or control effects in regression analyses. In addition, R&D-related investment shows substantial dispersion (SD = 6.6046), indicating considerable heterogeneity in firms’ innovation input intensity. Regarding innovation-related variables, both green patent applications and green patent grants exhibit highly skewed distributions, with means far below their maximum values, suggesting that a small number of firms account for a disproportionately large share of green innovation outputs. This pattern is consistent with the presence of innovation concentration among leading firms. For supply chain resilience indicators, the observed variation indicates differences in firms’ ability to diversify and stabilize supply chain relationships. Similarly, investment inefficiency measures show relatively small mean values but non-negligible dispersion, implying that misallocation of investment resources varies across firms. At the macro level, variables such as host-country GDP per capita, labor force participation rate, institutional quality, and trade openness display considerable variation, reflecting the diverse external environments faced by Chinese multinational enterprises. This heterogeneity is essential for identifying how firm-level ESG performance interacts with different institutional and market conditions.

3.4. Research Model

To examine the impact of ESG performance on firms’ OFDI resilience, we estimate the following baseline regression specifications using a fixed-effects panel model, with firm, year, and host-country fixed effects included in all specifications to control for unobserved heterogeneity.

Baseline model for resistance phase:

$${OFDIRes}_{\mathit{ijt}} = {\alpha }_0 + {\alpha }_1{\mathit{ESG}}_{\mathit{it}} + {\alpha }_2{\mathit{Controls}}_{\mathit{it}} + {\mu }_i +{ \lambda }_j + {\delta }_t +{ \epsilon }_{\mathit{ijt}}$$

(4)

Baseline model for recovery phase:

$${OFDIRec}_{\mathit{ijt}} ={ \beta }_0 +{ \beta }_1{\mathit{ESG}}_{\mathit{it}} +{ \beta }_2{\mathit{Controls}}_{\mathit{it}} + {\mu }_i + {\lambda }_j + {\delta }_t + {\epsilon }_{\mathit{ijt}}$$

(5)

In these specifications, i, j, and t index firms, host countries, and years, respectively. The dependent variables measure firms’ OFDI resilience in the resistance and recovery phases, while the key explanatory variable is firms’ ESG performance. The control variables include firm-level and host-country-level characteristics. All specifications include firm, year, and host-country fixed effects.

To test H2, H3, and H4, we construct a mediation model to examine the mediating roles of green innovation, supply chain resilience, and investment efficiency in the relationship between ESG performance and OFDI resilience [66].

The model is specified as follows:

$${Mechanism}_{\mathit{it}} = {\gamma }_0 +{ \gamma }_1{\mathit{ESG}}_{\mathit{it}} + {\gamma }_2{\mathit{Controls}}_{\mathit{it}} + {\mu }_i + {\lambda }_j + {\delta }_t + {\epsilon }_{\mathit{ijt}}$$

(6)

$${OFDIRes}_{\mathit{ijt}}={\theta }_0+{\theta }_1{\mathit{ESG}}_{\mathit{it}}+{{{\theta }_{2}Mechanism}_{\mathit{it}}+\theta }_3{\mathit{Controls}}_{\mathit{it}}+{\mu }_i+{\lambda }_j+{ \delta }_t+{ \epsilon }_{\mathit{ijt}}$$

(7)

$${OFDIRec}_{\mathit{ijt}}={\omega }_0+{\omega }_1{\mathit{ESG}}_{\mathit{it}}+{{ \omega }_{2}Mechanism}_{\mathit{it}}+{ \omega }_3{\mathit{Controls}}_{\mathit{it}}+{ \mu }_i+{ \lambda }_j+{\delta }_t+{\epsilon }_{\mathit{ijt}}$$

(8)

where ${Mechanism}_{\mathit{it}}$, represents the mediating mechanism, alternatively proxied by green innovation, supply chain resilience, and investment efficiency for firm i in year t. The remaining variables and fixed effects follow the baseline specification.

4. Empirical Analysis

4.1. Baseline Regression Results

Table 3. Regression results.

Variable	Resistance Phase			Recovery Phase
	(1)	(2)	(3)	(4)	(5)	(6)
	OFDIRes	OFDIRes	OFDIRes	OFDIRec	OFDIRec	OFDIRec
ESG	0.0145 ***	0.0174 ***	0.0181 ***	0.0293 ***	0.0313 ***	0.0431 ***
	(0.0050)	(0.0052)	(0.0054)	(0.0058)	(0.0075)	(0.0088)
Size		0.0494 ***	0.0524 ***		0.0189 ***	0.0175 ***
		(0.0151)	(0.0157)		(0.0047)	(0.0039)
ROE		0.0564 *	0.0508		0.0760 **	0.0796 **
		(0.0306)	(0.0319)		(0.0354)	(0.0351)
Top1		0.0010 **	0.0011 **		0.0005 ***	0.0006 ***
		(0.0004)	(0.0005)		(0.0001)	(0.0002)
SA		−0.2799 ***	−0.2901 ***		−0.5951 **	−0.6196 **
		(0.0955)	(0.1047)		(0.2817)	(0.3064)
Invent		0.0023	0.0024		0.0037	0.0036
		(0.0016)	(0.0016)		(0.0031)	(0.0034)
GDP			0.1048			0.0882 **
			(0.1105)			(0.0418)
LP			0.0073 *			0.0083 **
			(0.0044)			(0.0042)
WGI			0.0239 *			0.0943 *
			(0.0131)			(0.0557)
Trade			0.0012 *			0.0008 ***
			(0.0007)			(0.0001)
Constant	0.7178 ***	0.6317 **	1.5376 ***	1.4503 ***	1.1134 ***	2.7674 ***
	(0.1222)	(0.2959)	(0.3991)	(0.1623)	(0.3253)	(0.8228)
Id FE	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes
Country FE	Yes	Yes	Yes	Yes	Yes	Yes
R2_adjust	0.2324	0.2385	0.2465	0.3178	0.3635	0.3145
N	6849	6849	6849	12,872	12,872	12,872

Note: Standard errors in parentheses. *** p < 0.01. ** p < 0.05. * p < 0.1.

reports the baseline regression results examining the relationship between ESG performance and OFDI resilience across different phases of external shocks. All specifications include firm, year, and country fixed effects. Columns (1) to (3) present the results for the resistance phase, while columns (4) to (6) report the results for the recovery phase. Firm-level control variables are added in columns (2) and (5), and country-level controls are further included in columns (3) and (6).

ESG performance is positively and statistically significantly associated with OFDI resilience in the resistance phase. In column (3), the estimated coefficient on ESG performance is 0.0181 (p < 0.01), indicating that a one-unit increase in ESG performance is associated with a 0.0181 increase in OFDI resilience. This suggests that ESG engagement enhances firms’ ability to buffer external shocks and maintain the stability of overseas investment activities. The association is stronger in the recovery phase. In column (6), the coefficient increases to 0.0431 (p < 0.01), more than twice the magnitude observed in the resistance phase. This stronger effect implies that ESG performance plays a more prominent role in facilitating post-shock recovery, consistent with the argument that ESG-related capabilities contribute to rebuilding legitimacy, restoring stakeholder trust, and enabling strategic reconfiguration.

These findings provide direct empirical support for Hypotheses 1a and 1b, which predict that ESG performance positively affects OFDI resilience in both the resistance and recovery phases. Moreover, the larger coefficient in the recovery phase suggests a phase-dependent effect, highlighting the dynamic role of ESG across different stages of external shocks.

4.2. Endogeneity Discussion

4.2.1. Instrumental Variable

To address potential endogeneity concerns arising from omitted variables and reverse causality in estimating the association between ESG performance and OFDI resilience, an instrumental variable (IV) approach is used. We use the number of Pan-ESG funds holding shares in a focal firm in a given year as an instrument for firm-level ESG performance [67]. This IV satisfies the relevance condition, as pan-ESG funds can influence firms’ ESG performance by explicitly incorporating ESG criteria into their investment decisions. Firms with greater exposure to such funds are therefore more likely to exhibit stronger ESG performance.

The exogeneity of the instrument rests on the argument that the formation, scale, and portfolio allocation of pan-ESG funds are primarily driven by broader capital-market conditions and regulatory developments, rather than by firm-specific outcomes related to OFDI resilience. While participation by pan-ESG funds may affect firms’ ESG engagement, it is unlikely to directly influence firms’ OFDI resilience except through its impact on ESG performance. This institutional separation helps mitigate concerns that the instrument is directly correlated with unobserved determinants of OFDI resilience.

Table 4. Two-stage least squares and instrumental variable.

Variable	Resistance Phase		Recovery Phase
	(1)	(2)	(3)	(4)
	1st Stage	2nd Stage	1st Stage	2nd Stage
	ESG	OFDIRes	ESG	OFDIRec
IV	0.0639 ***		0.1155 ***
	(0.0203)		(0.0421)
ESG		0.1071 ***		0.1336 ***
		(0.0364)		(0.0305)
Controls	Yes	Yes	Yes	Yes
Id FE	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes
Country FE	Yes	Yes	Yes	Yes
KP rk Wald F	39.886		43.472
KP rk LM	12.638		13.947
	[0.0008]		[0.0009]
N	6849	6849	12,872	12,872

Note: Standard errors in parentheses. Data in square bracket are the p-values corresponding to the KP rk LM statistic. *** p < 0.01.

reports the results of the two-stage least squares (2SLS) estimations. The first-stage regressions show that the instrument is positively and significantly associated with ESG performance, indicating a strong correlation between the instrument and the endogenous regressor. The corresponding Kleibergen–Paap rk Wald F statistics exceed conventional thresholds, suggesting that weak-instrument concerns are unlikely. The Kleibergen–Paap rk LM statistics further reject the null hypothesis of underidentification. The second-stage results indicate that the estimated coefficients of ESG on OFDI resilience remain positive and statistically significant in both the resistance phase and the recovery phase. In economic terms, this suggests that improvements in ESG performance continue to enhance firms’ ability to stabilize and restore their overseas investment activities even after addressing potential endogeneity concerns.

Importantly, these findings provide further support for Hypotheses 1a and 1b, reinforcing the conclusion that ESG performance has a positive effect on OFDI resilience across different phases of external shocks.

4.2.2. Heckman Two-Step Method

Because ESG disclosure is largely voluntary, firms without ratings may be systematically excluded from the sample. To address potential sample selection bias, we implement the Heckman two-step model [68]. In the first stage, we estimate a Probit model predicting the likelihood that a firm receives an ESG rating, using the average ESG performance of other firms in the same industry (IndaESG) as the exclusion restriction. IndaESG captures industry-level ESG norms and peer effects that influence a firm’s decision to engage in ESG disclosure and obtain ratings. Conditional on firm, year, and country fixed effects as well as controls, IndaESG is unlikely to directly affect an individual firm’s OFDI resilience, satisfying the exclusion restriction. This is because industry-level ESG norms primarily influence disclosure behavior rather than firms’ operational resilience outcomes, ensuring the validity of the exclusion restriction.

The first-stage results are reported in columns (1) and (3) of

Table 5. Heckman two-stage robustness test.

Variable	Resistance Phase		Recovery Phase
	(1)	(2)	(3)	(4)
	1st Stage	2nd Stage	1st Stage	2nd Stage
	ESG	OFDIRes	ESG	OFDIRec
ESG		0.0176 **		0.0428 **
		(2.395)		(2.277)
IMR		−0.0087		−0.0295
		(−0.253)		(−0.581)
IndaESG	0.6222 ***		1.0069 ***
	(12.182)		(12.408)
Controls	Yes	Yes	Yes	Yes
Id FE	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes
Country FE	Yes	Yes	Yes	Yes
N	6849	6849	12,872	12,872

Note: Standard errors in parentheses. *** p < 0.01. ** p < 0.05.

. They show that IndaESG is positively and highly significantly associated with firm-level ESG performance. In the second stage, the inverse Mills ratio (IMR) derived from the first stage is included in the baseline regression. As shown in columns (2) and (4), the coefficients on ESG remain positive and statistically significant for OFDI resilience in both the resistance phase and the recovery phase. The IMR is statistically insignificant in both specifications, indicating that sample selection bias is unlikely to drive the main results. This suggests that the estimated relationship is not systematically biased by the non-random selection of firms with ESG ratings.

Importantly, these findings provide further support for Hypotheses 1a and 1b, reinforcing the conclusion that ESG performance positively affects OFDI resilience across both phases.

4.3. Robustness Test

4.3.1. Replacement of the Core Independent Variable

To examine whether our baseline results are sensitive to different measures of ESG performance, we employ two additional approaches. First, we replace the CNRDS ESG score with Bloomberg ESG scores, which adopt a distinct global methodology that places greater emphasis on disclosure quality and international comparability [69]. This alternative measure provides a more internationally comparable benchmark and helps alleviate concerns that the baseline results are driven by a specific domestic ESG rating system. Second, we transform the continuous CNRDS ESG score into a three-category ordinal variable based on yearly terciles, assigning values of 3, 2, and 1 to firms in the top, middle, and bottom terciles, respectively [70]. This specification captures the relative ranking of firms’ ESG performance and allows us to test whether the results are robust to potential nonlinearities and ordinal scaling.

The results reported in columns (1) to (4) of

Table 6. Replacement of core independent variable and dependent variable.

Variable	Replacement of Core Independent Variable				Dependent Variable
	Resistance Phase		Recovery Phase
	(1)	(2)	(3)	(4)	(5)
	OFDIRes	OFDIRes	OFDIRes	OFDIRec	Resilience
Bloomberg ESG	0.0145 **	0.0289 ***
	(0.0065)	(0.0073)
ESG tercile rank			0.0035 ***	0.0047 ***
			(0.0013)	(0.0011)
ESG					0.0286 ***
					(0.0051)
Constant	0.0391	3.2089 **	3.0374 ***	10.2630 ***	−1.6449
	(0.0316)	(1.5289)	(1.0984)	(3.2030)	(2.4413)
Controls	Yes	Yes	Yes	Yes	Yes
Id FE	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes
Country FE	Yes	Yes	Yes	Yes	Yes
R2_adjust	0.2630	0.2134	0.2632	0.1938	0.3172
N	6849	6849	12,872	12,872	12,872

Note: Standard errors in parentheses. *** p < 0.01. ** p < 0.05.

remain qualitatively consistent with the baseline findings. The coefficients on both the Bloomberg ESG score and the ESG tercile indicator are positive and statistically significant for OFDI resilience in both the resistance phase and the recovery phase. In economic terms, this suggests that improvements in ESG performance—regardless of how they are measured—consistently enhance firms’ ability to withstand external shocks and recover from disruptions in overseas investment activities.

Importantly, these findings provide further support for Hypotheses 1a and 1b, indicating that the positive relationship between ESG performance and OFDI resilience is robust to alternative data sources and measurement specifications.

4.3.2. Replacement of the Dependent Variable

Following Han et al. (2025) [10], we construct an integrated measure of OFDI resilience (Resilience) that aggregates performance across the entire shock episode without distinguishing between the resistance and recovery phases. Specifically, Han et al. (2025) measure OFDI resilience as the deviation between a firm’s actual change in overseas investment performance and the expected change in the absence of shocks, where the expected change is estimated based on the firm’s overall OFDI growth trend. This alternative measure differs from our phase-specific measures in one fundamental respect: whereas our measures decompose resilience into a resistance phase and a recovery phase to capture firms’ distinct capacities to absorb shocks and to reconfigure performance thereafter, this integrated index aggregates performance across the entire shock episode without distinguishing between the two stages. This alternative specification provides a broader assessment of firms’ overall resilience and helps verify that the baseline results are not sensitive to the phase-based decomposition. The regression results using this alternative dependent variable are reported in column (5) of Table 6. ESG performance remains positively and statistically significantly associated with overall OFDI resilience. This indicates that ESG performance contributes not only to phase-specific resilience but also to firms’ overall ability to manage and recover from external shocks. These results further confirm Hypotheses 1a and 1b and demonstrate that the main conclusions are robust to alternative operationalizations of OFDI resilience.

4.3.3. Further Robustness Tests

We conduct three complementary sets of robustness checks to examine the reliability of our baseline findings. First, we refine the sample by excluding observations during China’s 2015–2016 stock market crash and investments in host countries with armed conflict or severe political instability (Ukraine, Azerbaijan, Nigeria, Uganda, Kenya, the Democratic Republic of the Congo, Liberia, and Zambia). This restriction ensures that the estimated relationship is not driven by extreme macroeconomic shocks or atypical investment environments. Second, we winsorize dependent variables at the 0.5% level in both tails to mitigate the influence of extreme values, which reduces the sensitivity of the estimates to outliers while preserving the overall distribution of the data [71]. This approach ensures that the results are not disproportionately affected by a small number of extreme observations. Third, given that both dependent variables contain a substantial proportion of zeros, we re-estimate the baseline model using a high-dimensional fixed effects Poisson pseudo-maximum likelihood estimator (PPMLHDFE) [72].

The results are reported in

Table 7. Other robustness tests.

Variable	Refining the Sample		Winsorization		Alternative Estimation
	(1)	(2)	(3)	(4)	(5)	(6)
	OFDIRes	OFDIRec	OFDIRes	OFDIRec	OFDIRes	OFDIRec
ESG	0.0136 ***	0.0261 ***	0.0174 **	0.0426 **	0.0143 ***	0.0249 ***
	(0.0031)	(0.0024)	(0.0073)	(0.0188)	(0.0042)	(0.0016)
Constant	0.6218 ***	0.8576 ***	−0.2125 ***	2.7734 ***	0.7984	−1.7352 **
	(0.1452)	(0.1597)	(0.0372)	(0.8319)	(0.7259)	(0.6767)
Controls	Yes	Yes	Yes	Yes	Yes	Yes
Id FE	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes
Country FE	Yes	Yes	Yes	Yes	Yes	Yes
R2_adjust	0.2448	0.2692	0.2623	0.2148	0.3091	0.4241
N	5892	11,447	6849	12,872	6849	12,872

Note: Standard errors in parentheses. *** p < 0.01. ** p < 0.05.

. Columns (1) and (2) use the refined sample, columns (3) and (4) apply winsorization, and columns (5) and (6) report the PPMLHDFE results. ESG performance remains positively and statistically significantly associated with OFDI resilience. These findings provide further support for Hypotheses 1a and 1b, confirming that the positive relationship between ESG performance and OFDI resilience is robust and not driven by specific sample conditions, outliers, or model specifications.

4.4. Mediation Results

4.4.1. Green Innovation as a Mediating Channel

Following the two-step procedure of Hayes (2017) for mediation analysis [66], we test whether green innovation serves as a mediating channel through which ESG performance enhances OFDI resilience. The results in columns (1) and (2) of

Table 8. Mediating effect of green innovation.

Variable	Resistance Phase				Recovery Phase
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	GPA	GPG	OFDIRes	OFDIRes	GPA	GPG	OFDIRec	OFDIRec
ESG	0.7591 **	0.1441 ***	0.0161 **	0.0164 **	1.5869 *	0.9208 **	0.0415 **	0.0411 **
	(0.2957)	(0.0475)	(0.0072)	(0.0071)	(0.9057)	(0.3853)	(0.0188)	(0.0189)
GPA			0.0026 ***				0.0010 **
			(0.0005)				(0.0004)
GPG				0.0118 ***				0.0022 ***
				(0.0005)				(0.0006)
Constant	196.5598 ***	−9.5 × 102 ***	0.6868	−0.1129 ***	73.4747	26.1055	3.7492 ***	2.7791 ***
	(75.6360)	(279.0841)	(0.0275)	(0.0269)	(52.1614)	(176.7356)	(0.8798)	(0.8251)
Controls	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Id FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Country FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
R2_adjust	0.5104	0.4206	0.5634	0.4198	0.4622	0.4970	0.4948	0.5326
N	6849	6849	6849	6849	12,872	12,872	12,872	12,872

Note: Standard errors in parentheses. *** p < 0.01. ** p < 0.05. * p < 0.1.

show that ESG performance is positively and significantly associated with both green patent applications (GPA) and green patent grants (GPG) in the resistance phase. The coefficient on ESG decreases but remains significant after including green innovation variables, indicating a partial mediating effect. Importantly, these findings provide empirical support for the resource-based view proposed in the theoretical framework. In this stage, green innovation functions as a firm-specific strategic resource that enhances resource efficiency, reduces dependence on conventional inputs, and improves the stability of critical production factors [73]. As a result, firms are better able to buffer external shocks and maintain operational continuity, thereby strengthening OFDI resilience.

In the recovery phase, ESG performance continues to exert a positive and significant effect on both GPA and GPG, as shown in columns (5) and (6) of Table 8. In columns (7) and (8), both GPA and GPG significantly enhance OFDI recovery, while the ESG coefficient decreases but remains significant, again indicating partial mediation. Beyond statistical significance, the results also support the dynamic capability view underlying the recovery process. In this stage, green innovation enables firms to develop new technologies and environmentally aligned products, which facilitate adaptation to evolving host-country regulations and market conditions [74]. This allows firms to reconfigure their resource base, overcome institutional constraints, and pursue new growth opportunities, thereby supporting both recovery and long-term resilience of OFDI.

Taken together, the empirical findings confirm that green innovation is not merely correlated with ESG performance and OFDI resilience, but serves as a key transmission channel through which ESG enhances resilience across different stages. Its role evolves from a resource-based buffering mechanism in the resistance phase to a capability-driven adaptation mechanism in the recovery phase. Therefore, Hypotheses 2a and 2b are supported.

4.4.2. Supply Chain Resilience as a Mediating Channel

Table 9. Mediating effect of supply chain resilience.

Variable	Resistance Phase				Recovery Phase
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	SCC	SCR	OFDIRes	OFDIRes	SCC	SCR	OFDIRec	OFDIRec
ESG	−0.4593 ***	−0.0209 ***	0.0158 **	0.0169 **	−0.2166	0.0038	0.0436 **	0.0448 **
	(0.1204)	(0.0059)	(0.0074)	(0.0074)	(0.1679)	(0.0084)	(0.0188)	(0.0189)
SCC			−0.0050 ***				−0.0016
			(0.0008)				(0.0017)
SCR				−0.0574 ***				0.0350
				(0.0186)				(0.0386)
Constant	100.1735 ***	−8.0613 ***	1.1902 ***	−1.5631 ***	−29.7935	−8.0269 ***	1.7358 **	2.2438 **
	(35.9822)	(1.5673)	(0.0477)	(0.1108)	(68.0106)	(3.0572)	(0.8274)	(0.8573)
Controls	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Id FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Country FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
R2_adjust	0.4197	0.5411	0.4663	0.59978	0.4693	0.5471	0.4112	0.5643
N	6849	6849	6849	6849	12,872	12,872	12,872	12,872

Note: Standard errors in parentheses. *** p < 0.01. ** p < 0.05.

reports the regression results for the relationship between ESG performance and two dimensions of supply chain resilience, namely supply chain concentration (SCC) and supply chain resistance (SCR). As shown in columns (1) and (2), ESG performance is negatively and significantly associated with both SCC and SCR. In columns (3) and (4), both SCC and SCR exert significant negative effects on OFDI resilience, while the coefficient on ESG decreases in magnitude but remains statistically significant after including the mediators. In the resistance phase, the results provide strong support for the mediating effect of supply chain resilience from a resource-based perspective. This is because ESG performance facilitates the formation of diversified and trust-based supplier networks, which serve as valuable firm-specific resources that reduce dependence on single suppliers, enhance coordination efficiency, and enable firms to absorb shocks and maintain operational continuity under external disruptions [75]. Therefore, Hypothesis 3a is supported.

In the recovery phase, the results provide no support for the mediating effect of supply chain resilience from a resource-based perspective. ESG performance has no significant effect on either SCC or SCR. A possible explanation is that managerial attention shifts toward strategic reconfiguration, resource redeployment, and opportunity exploration [76]. As a result, although ESG performance continues to enhance OFDI resilience, its influence is no longer primarily transmitted through supply chain resilience. Instead, the supply chain channel becomes less salient, as it is overshadowed by competing strategic priorities, thereby weakening its mediating role and rendering it statistically insignificant. Therefore, Hypothesis 3b is not supported.

4.4.3. Investment Efficiency as a Mediating Channel

In this section, we test whether investment efficiency serves as a mediating channel through which ESG performance enhances OFDI resilience. We examine two established measures of investment efficiency (IE1 and IE2). Columns (1) and (2) in

Table 10. Mediating effect of investment efficiency.

Variable	Resistance Phase				Recovery Phase
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	IE1	IE2	OFDIRes	OFDIRes	IE1	IE2	OFDIRec	OFDIRec
ESG	0.0023 ***	0.0013 **	0.0137 ***	0.0113 ***	0.0010	0.0008	0.0401 **	0.0418 **
	(0.0005)	(0.0006)	(0.0015)	(0.0106)	(0.0010)	(0.0011)	(0.0186)	(0.0178)
IE1			1.9130 ***				0.5665
			(0.3347)				(0.5121)
IE2				5.2308 ***				1.0602
				(0.2809)				(0.8306)
Constant	−4.7675	43.3445 **	1.0380	2.2942 ***	−2.0461 **	9.9987 *	2.7141 ***	1.6370 **
	(4.9101)	(17.8712)	(0.7291)	(0.7021)	(0.9447)	(5.4107)	(0.0184)	(0.7164)
Controls	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Id FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Country FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
R2_adjust	0.4371	0.4790	0.4657	0.4742	0.5184	0.5952	0.5151	0.5692
N	6849	6849	6849	6849	12,872	12,872	12,872	12,872

Note: Standard errors in parentheses. *** p < 0.01. ** p < 0.05. * p < 0.1.

report the regression results for the resistance phase. ESG performance is positively and significantly associated with both IE1 and IE2. In columns (3) and (4), both IE1 and IE2 exert significant positive effects on OFDI resilience, while the coefficient of ESG decreases but remains statistically significant after including the mediators, indicating a partial mediating effect. These findings suggest that, in the resistance phase, the investment efficiency gains derived from strong ESG performance operate as a firm-specific strategic asset consistent with the RBV. Stronger internal governance mechanisms and closer alignment between managerial incentives and long-term shareholder value promote more disciplined and precise capital allocation decisions, curtailing unnecessary expenditures and directing financial resources toward high-priority operations [77]. Therefore, Hypothesis 4a is supported.

In the recovery phase (columns (5) to (8)), the mediating role of investment efficiency becomes insignificant. ESG performance has no significant effect on either IE1 or IE2. In columns (7) and (8), neither IE1 nor IE2 exerts a significant influence on OFDI resilience in the recovery phase, while the coefficient on ESG remains positive and significant. One possible reason is that managerial pressure to restore operational stability may induce overinvestment in short-term crisis response and reputational rehabilitation, leading to resource misallocation that temporarily disrupts the governance discipline through which ESG performance normally enhances investment efficiency [78]. Therefore, Hypothesis 4b is not supported.

4.5. Heterogeneity Analyses

4.5.1. Industry Environmental Sensitivity

Prior research suggests that industry-level environmental sensitivity shapes firms’ ESG engagement and its strategic implications in international markets [79]. In particular, environmental sensitivity fundamentally shapes firms’ motivation and capacity to engage in green innovation, which is a central mechanism through which ESG performance affects OFDI resilience. Following this logic, we partition the sample into high- and low-environmental-sensitivity industries based on whether firms operate in heavily polluting sectors. As shown in

Table 11. Heterogeneity analysis by industry environmental sensitivity.

Variable	Resistance Phase		Recovery Phase
	High Sensitivity	Low Sensitivity	High Sensitivity	Low Sensitivity
	(1)	(2)	(3)	(4)
	OFDIRes	OFDIRes	OFDIRec	OFDIRec
ESG	0.0037	0.0209 ***	0.0546 ***	0.0377
	(0.0150)	(0.0081)	(0.0209)	(0.0237)
Constant	−3.1876 **	0.5320 *	−4.3513 *	3.7947 **
	(1.3087)	(0.2771)	(2.3002)	(1.5417)
Controls	Yes	Yes	Yes	Yes
Id FE	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes
Country FE	Yes	Yes	Yes	Yes
R2_adjust	0.2673	0.2655	0.2848	0.2011
N	1832	5017	3584	9288
p_value	0.0015		0.0028

Note: Standard errors in parentheses. *** p < 0.01. ** p < 0.05. * p < 0.1. The p-values are the results of the Fisher test.

, the relationship between ESG performance and OFDI resilience exhibits a clear phase-dependent pattern. In the resistance phase, ESG performance is positive but statistically insignificant for firms in high-environmental-sensitivity industries. In contrast, the coefficient is positive and statistically significant at the 1% level for firms in low-environmental-sensitivity industries. This pattern reverses in the recovery phase. ESG performance is strongly and positively associated with OFDI resilience for firms in high-environmental-sensitivity industries, whereas the effect becomes statistically insignificant for firms in low-environmental-sensitivity industries.

This phase-contingent pattern can be understood through differences in attribution and signal credibility across stages. During the resistance phase, heightened uncertainty and risk perceptions in host countries induce stakeholders to rely on heuristic-based attribution. Environmentally sensitive firms, which are already associated with pollution-intensive activities, are more likely to be targeted by eco-nationalism and blamed for negative shocks [80]. This attribution process is largely path-dependent and anchored in prior perceptions rather than updated ESG performance. As a result, ESG engagement—primarily perceived as a soft signal—fails to effectively alter stakeholder beliefs in the short run. The presence of greenwashing suspicion further weakens its credibility. Under such conditions, the mediating channels identified in this study, particularly green innovation, are not immediately recognized or valued, thereby limiting their ability to enhance OFDI resilience in the resistance phase. In the recovery phase, however, stakeholder evaluation gradually shifts from heuristic-based judgment to evidence-based assessment. Substantive green innovation outputs, such as green patents, provide verifiable and difficult-to-mimic signals of firms’ environmental commitment and technological capability [81].

For firms in low-environmental-sensitivity industries, during the resistance phase, green innovation plays a significant mediating role primarily through an efficiency-enhancing mechanism. Unlike firms in environmentally sensitive industries, these firms are not subject to strong negative prior perceptions or stigma. As a result, their ESG engagement and related green innovation activities are less likely to be interpreted as symbolic or opportunistic behavior. Instead, green innovation can be directly translated into operational improvements, such as enhanced resource efficiency, reduced input dependence, and more stable production processes. These efficiency gains are particularly valuable under crisis conditions, where firms need to maintain continuity and absorb external shocks. Consequently, green innovation serves as an effective transmission channel through which ESG performance enhances OFDI resilience in the resistance phase. In the recovery phase, however, the mediating role of green innovation becomes insignificant. This can be explained by a shift in firms’ strategic priorities and the declining marginal contribution of green innovation. Firms in low-environmental-sensitivity industries typically face fewer legitimacy constraints and are less exposed to stakeholder skepticism [82]. Therefore, they do not need to rely on green innovation as a signaling device to restore trust or legitimacy. At the same time, the recovery phase is characterized by a greater emphasis on market expansion, strategic reconfiguration, and opportunity exploration, where the incremental benefits of green innovation are relatively limited in the short run. As a result, the marginal effect of green innovation diminishes, weakening its role as a mediating channel between ESG performance and OFDI resilience.

4.5.2. Degree of Digital Transformation

Digital transformation shapes firms’ information-processing capacity, risk management practices, and coordination efficiency in cross-border operations, which may condition how ESG performance translates into OFDI resilience [83]. We therefore divide the sample into high- and low-digital-transformation groups using the median value as the cutoff point.

Table 12. Heterogeneity analysis by degree of digital transformation.

Variable	Resistance Phase		Recovery Phase
	High Degree	Low Degree	High Degree	Low Degree
	(1)	(2)	(3)	(4)
	OFDIRes	OFDIRes	OFDIRec	OFDIRec
ESG	0.0205 **	0.0071	0.0465	0.0514 ***
	(0.0094)	(0.0122)	(0.0334)	(0.0105)
Constant	0.9216 *	−1.7757 ***	4.2652 ***	4.2873 **
	(0.4953)	(0.5570)	(1.2275)	(1.9307)
Controls	Yes	Yes	Yes	Yes
Id FE	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes
Country FE	Yes	Yes	Yes	Yes
R2_adjust	0.2738	0.2745	0.1964	0.1898
N	4427	2422	6234	6638
p_value	0.0301		0.0015

Note: Standard errors in parentheses *** p < 0.01 ** p < 0.05 * p < 0.1.

reveals a clear pattern that is both phase-dependent and asymmetric. ESG performance enhances OFDI resilience for high-digital-transformation firms during the resistance phase, but for low-digital-transformation firms during the recovery phase.

In the resistance phase, high-digital-transformation firms are better positioned to leverage ESG performance as a resilience-enhancing mechanism because digital transformation functions as a complementary asset that activates and amplifies the value of ESG-related resources. Under the RBV, technologies like predictive analytics, platform ecosystems, and real-time data systems enhance operational coordination, optimize resource utilization, and embed ESG considerations into value-creating processes. This complementarity ensures that ESG-related assets can be rapidly converted into defensive actions when sudden disruptions strike host-country operations. Such defensive actions demand speed, which is precisely what digital infrastructure provides. Low-digital-transformation firms, lacking this complementary infrastructure, cannot activate their ESG assets at the pace that the resistance phase demands, leaving ESG performance unable to translate into meaningful OFDI resilience gains during this phase.

In the recovery phase, this complementarity logic inverts. Digital transformation shifts from an enabler of ESG to a substitute for it. Recovery demands active reconfiguration: sensing emerging opportunities, restoring stakeholder legitimacy, and redirecting resources toward new growth trajectories. For high-digital-transformation firms, these reconfiguration demands are increasingly met by digital dynamic capabilities rather than ESG mechanisms. As digital capabilities assume the reconfiguration functions previously performed by ESG, the incremental contribution of ESG to OFDI resilience diminishes, rendering its effect statistically insignificant for high-digital-transformation firms during recovery [84]. For low-digital-transformation firms, no such substitution occurs. Absent the digital infrastructure to drive recovery independently, these firms remain reliant on ESG performance as their primary reconfiguration mechanism. Through this reliance, they rebuild stakeholder trust, restore institutional legitimacy, and realign governance structures with post-shock host-country requirements. This uncontested reliance allows ESG performance to translate into OFDI resilience with full force, producing the significant effect observed in this group during the recovery phase.

4.5.3. Host-Country Income Level

Host-country institutional and market environments vary systematically with income levels, potentially shaping the effectiveness of ESG performance in enhancing OFDI resilience. We classify host countries into high-income (HI) versus middle- and low-income (LMI) groups based on World Bank income classifications [85]. As shown in

Table 13. Heterogeneity tests by host-country income level.

Variable	Resistance Phase		Recovery Phase
	HI	LMI	HI	LMI
	(1)	(2)	(3)	(4)
	OFDIRes	OFDIRes	OFDIRec	OFDIRec
ESG	0.0105	0.0255 **	0.0502	0.0477 **
	(0.0092)	(0.0115)	(0.0298)	(0.0205)
Constant	−0.5470 ***	0.3906	0.2399	0.2853
	(0.1727)	(0.7073)	(0.5026)	(0.5280)
Controls	Yes	Yes	Yes	Yes
Id FE	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes
Country FE	Yes	Yes	Yes	Yes
R2_adjust	0.2537	0.2532	0.1509	0.265
N	3967	2882	7291	5581
p_value	0.0352		0.0229

Note: Standard errors in parentheses. *** p < 0.01. ** p < 0.05. The p-values are the results of the Fisher test.

, the association between ESG performance and OFDI resilience varies markedly across host-country income groups. ESG performance is positively and statistically significantly associated with OFDI resilience in both the resistance and recovery phases for middle- and low-income economies. The corresponding coefficients for high-income economies are statistically insignificant in both phases. These results suggest that the effectiveness of ESG performance is contingent on host-country institutional environments.

High-income economies have well-established institutional frameworks, strong rule of law, stringent ESG regulations, and well-functioning market mechanisms that provide robust safeguards and stakeholder protections. As a result, additional ESG investments yield limited incremental benefits, leading to statistically insignificant associations with OFDI resilience in both the resistance and recovery phases. In such contexts, formal institutions substitute for firm-level ESG efforts, thereby weakening the marginal contribution of ESG performance. In middle- and low-income economies, formal institutions are weaker and institutional voids are more pronounced. ESG performance functions as an important informal signal of trustworthiness and legitimacy. This signal can shape government tolerance, social acceptance, and reputational capital, generating stronger legitimacy benefits and policy support during both crisis and recovery periods. As a result, ESG practices are more likely to translate into enhanced OFDI resilience in these contexts compared to high-income economies.

5. Discussion

5.1. Main Conclusions

This study examines Chinese multinational enterprises listed on the A-share market from 2008 to 2024, focusing on how ESG performance shapes OFDI resilience across distinct post-shock periods, the underlying mediating mechanisms, and heterogeneity across firm and industry contexts. The main conclusions are as follows:

ESG performance enhances OFDI resilience across both phases. ESG performance is positively and significantly associated with stronger OFDI resilience throughout the disruption cycle, spanning both the resistance and recovery phases. This relationship remains stable after robustness checks and endogeneity corrections, consistent with emerging evidence that ESG practices function as strategic assets that extend beyond regulatory compliance to support firms’ capacity to absorb and navigate external disruptions in international markets.

The mediating channels through which ESG enhances OFDI resilience vary across phases. Green innovation serves as a consistent transmission channel in both the resistance and recovery phases, making it the most stable mediating mechanism linking ESG performance to OFDI resilience throughout the disruption cycle. By contrast, supply chain resilience and investment efficiency function as significant mediating channels only in the resistance phase and become statistically insignificant in the recovery phase.

These findings speak to and extend two adjacent streams of literature. First, they extend the literature on ESG and firm resilience. Prior studies have established that ESG practices enhance firm resilience predominantly in domestic market contexts [86]. Our study extends this line of inquiry to OFDI, demonstrating that the resilience-enhancing function of ESG operates across firms’ international investments, which are subject to greater institutional complexity, political risk, and operational uncertainty. Second, they contribute to the nascent literature on OFDI resilience. Existing studies on OFDI resilience remain limited, with prior work focusing primarily on the role of artificial intelligence in enabling firms to sustain and adapt their foreign investment under adverse conditions [10]. Our study disaggregates OFDI resilience into distinct resistance and recovery phases, revealing that the drivers and mechanisms of resilience differ systematically across these stages.

Overall, these findings enrich the empirical literature on ESG and international business resilience by demonstrating that the value of ESG is phase-dependent, functioning as a static resource buffer under the RBV during resistance and as a dynamic reconfiguration mechanism under the DCV during recovery, and is also context-contingent, varying systematically with industry characteristics, digital capabilities, and host-country institutional environments.

5.2. Practical Implications

Based on the empirical findings, this study offers several practical implications for emerging-market multinationals, investors, and policymakers. A central finding is that ESG should be viewed not as a compliance cost, but as a strategic asset whose effectiveness varies across disruption phases and institutional contexts.

For emerging-market multinational enterprises, ESG considerations should be integrated into core OFDI strategies rather than treated as peripheral compliance costs. Firms operating in volatile host-country environments may benefit from investing in green innovation and supply chain diversification before disruptions materialize, as these assets function most effectively when pre-accumulated and can be immediately deployed under acute time constraints. In particular, green innovation—identified as the most consistent mediating channel—should be prioritized as a long-term strategic capability.

For investors and capital providers, investment decisions should extend beyond short-term financial returns to incorporate assessments of firms’ ESG-driven resilience potential. Firms with strong ESG performance, complementary digital capabilities, and established green innovation capacity represent lower-risk OFDI profile, particularly in volatile emerging markets and institutionally weak host-country environments. Given that green innovation operates across both resistance and recovery phases, it serves as a particularly reliable indicator of long-term resilience.

For policymakers, both home- and host-country governments can play an active role in strengthening the ESG–OFDI resilience relationship. Home-country policies that incentivize green innovation investment and digital transformation—through tax incentives, R&D subsidies, and ESG disclosure standards—can enhance the resilience capacity of outward investors before they enter volatile markets. Host-country governments in middle- and low-income economies may attract more stable and resilient foreign investment by strengthening institutional alignment with international ESG standards, reducing legitimacy barriers, and rewarding substantive ESG engagement rather than symbolic compliance.

5.3. Research Limitations and Future Research Directions

Although mediating mechanisms and heterogeneity effects are examined separately, the theoretical mechanisms through which digital transformation level and host-country income level condition the transmission efficiency of individual mediating channels remain insufficiently articulated, leaving the internal theoretical relationship between heterogeneity results and mediating pathways in need of further conceptual development. This theoretical gap is compounded by a measurement limitation: treating ESG performance as an annual cross-sectional indicator may not fully capture the dynamic and cumulative nature of ESG practices over time, as the accumulation effects and time-lagged impacts of ESG investment on OFDI resilience may follow nonlinear trajectories that a static measure cannot adequately reflect. A further boundary condition concerns shock heterogeneity—this study integrates the global financial crisis, US–China trade tensions, and the COVID-19 pandemic into a unified analytical framework without distinguishing among their fundamentally different transmission mechanisms, durations, and scope of impact, leaving open the question of whether the phase-dependent ESG–OFDI resilience relationship varies systematically across financial, geopolitical, and public health shocks.

These limitations point toward directions for future research. More integrative theoretical frameworks are needed to specify how digital capabilities and host-country institutional environments shape the transmission efficiency of individual mediating channels across resilience phases. Dynamic panel models or event-study designs could better capture the cumulative and time-lagged effects of ESG investment on resilience outcomes. Finally, shock-specific research designs would further clarify whether ESG functions as a more effective resilience mechanism under financial, geopolitical, or public health disruptions.