The Synergistic Trap: How Strategic Alliances Amplify Corporate Vulnerability to Climate Risk

Abstract

As climate change increasingly challenges corporate operations and sustainable development, the role of strategic alliances in managing environmental risks requires critical reassessment. While prior research highlights their benefits for innovation and performance, potential adverse consequences in the face of climate risks remain underexplored. Using panel data of Chinese A-share listed firms from 2010 to 2023, this study applies econometric models to evaluate the impact of strategic alliances on firms’ climate risk exposure. The findings show that strategic alliances significantly weaken firms’ resilience to climate risks by diverting executive attention from environmental issues, constraining sustainability capacity building, and reducing sensitivity to supply chain risks. These adverse effects are more pronounced for firms with poor carbon performance and lower firm value. Moreover, compared with contractual alliances, equity-based alliances create deeper binding and reduce flexibility in responding to climate change. The study contributes to theory and practice by suggesting that firms should optimize alliance structures, increase partner heterogeneity, and enhance executive awareness of climate risks to improve resilience in the context of climate governance.

1. Introduction

In recent years, the severity of climate change has expanded both in scope and intensity. It transforms the business environment in unprecedented ways. According to IPCC [1] and TCFD [2], climate risk exposure includes not only extreme weather events and natural disasters. It also involves operational challenges such as supply chain disruptions, rising energy costs, and stricter carbon regulations. Beyond these physical and transition risks, global initiatives such as the Paris Agreement, the European Green Deal, and China’s “dual-carbon goals” further compel firms to integrate climate considerations into strategic decision-making. Increasing investor demands for ESG disclosure and mounting stakeholder pressure also highlight that climate risk management has become not only a social responsibility but also a core element of long-term competitiveness. Therefore, it is imperative for corporations to strategically incorporate climate risk management to improve environmental resilience and maintain competitive advantages.

Strategic alliances can help firms mitigate climate-related risks by enhancing interfirm collaboration [3], supporting green technology development [4], and reducing information barriers [5]. Strategic alliances have also been increasingly recognized in the context of sustainability, for instance in green supply chain management, knowledge sharing for clean technologies, and cross-industry partnerships that accelerate the transition to low-carbon economies. However, most existing studies focus on the positive effects of strategic alliances on economic performance and innovation efficiency, while insufficient attention has been paid to their potential environmental downsides.

Existing studies often assume alliances act as a buffer to distribute risks but rarely explore when they might instead amplify vulnerabilities. For example, differences in environmental capabilities and commitment across partners, coupled with complex governance structures, may weaken the effectiveness of green strategies within alliances. This gap is particularly critical because overlooking the possible negative consequences of alliances may lead firms to underestimate their true exposure to climate risks. Hence, studying the role of strategic alliances in climate risk management not only enriches the theoretical framework of environmental management and organizational collaboration, but also provides practical strategic guidance for firms in responding to climate risks.

This study systematically examines the impact of strategic alliances in climate risk management. In contrast to existing literature that views alliances as a strategic response to climate risk, this paper introduces a new viewpoint: in certain contexts, strategic alliances may exacerbate rather than mitigate firms’ climate vulnerability. In summary, this study contributes to the literature in three important ways: (1) Measurement innovation: It develops a text-based index to capture corporate climate risk exposure, enriching existing approaches to climate risk assessment; (2) Mechanism analysis: It proposes and empirically tests three mediating mechanisms—executive attention diversion, sustainability capability crowding-out, and supply chain risk underestimation—through which strategic alliances influence climate risk; (3) Heterogeneity exploration: It examines how the impacts of alliances differ depending on carbon performance, firm value, and alliance structure, providing nuanced insights into when alliances mitigate versus exacerbate climate risks.

2. Literature Review

2.1. Related Literature on Climate Risks

Amid worsening climate conditions worldwide, climate risk is now recognized as a critical form of non-conventional risk for businesses. Climate risk, as outlined by the IPCC [6] and TCFD [2], involves the uncertain consequences of climate change that can influence a corporation’s operations, asset value, financial standing, and strategic decisions. Typically, climate risks are categorized into two types: physical risks and transition risks [7]. Physical risks arise from disruptions to infrastructure and supply chains caused by extreme weather events or long-term climate change. Transition risks, on the other hand, are associated with policy shifts, institutional changes, technological innovations, and changes in market behavior in response to climate change.

In contrast to conventional risks, climate risk differs markedly in terms of its causation, extent of influence, and the degree to which it can be anticipated. Traditional risks, driven by market fluctuations or internal managerial errors, are typically characterized by a certain degree of dispersion and relatively high predictability [8,9]. Climate risk, in contrast, is marked by high degrees of system-wide relevance and external spillovers. Its consequences are extensive and can diffuse swiftly through financial and economic linkages [10]. Another distinctive feature of climate risk is its time-based variability. It can produce prolonged effects and trigger rapid escalation when crucial tipping points are reached [11]. These features challenge the effectiveness of conventional modeling frameworks in capturing climate risk. Consequently, the systemic complexity and unpredictability of climate processes only heighten the difficulty of developing reliable models and metrics.

Conventional static approaches are no longer sufficient for managing climate risk. A more flexible and integrated framework is needed—one that emphasizes coordination across systems, time horizons, and fields of expertise [12]. Currently, tools like scenario planning [13], risk materiality matrices [14], and carbon-linked financial metrics [15] have become common in climate risk management. Their use has steadily deepened and broadened how organizations identify and evaluate climate-related risks. Importantly, a hybrid use of qualitative and quantitative methods is gaining traction in climate risk assessment. Techniques such as expert assessments, scenario-based climate modeling, and DSGE frameworks are being adopted to address issues of data scarcity and uncertainty in projections [16,17].

At the corporate level, firms need to mitigate and gradually adapt to climate risk. On one side, reducing emissions and increasing green investments can help lower the probability of future climate risks. On the other, firms must improve emergency preparedness and strengthen infrastructure resilience to better cope with unavoidable disruptions [1,18].

In general, climate risk has increasingly become a central concern in macroeconomics, financial governance, and corporate strategic planning [19]. For example, investors are increasingly factoring climate considerations into asset valuation to mitigate return uncertainties linked to high-carbon assets [7]. Additionally, firms respond by improving financial robustness through restructuring their capital composition [20], whereas governments play a guiding role by fostering green innovation and upgrading industries through supportive policies and incentives [21]. Clearly, research on climate risk is evolving into a crucial interdisciplinary domain that integrates environmental science with economic and managerial studies. Future progress will require continued efforts in both theory and practice.

2.2. Related Literature on Strategic Alliances

Strategic alliances involve cooperation between firms through means such as agreements or joint ventures, allowing them to share resources, expertise, and markets while remaining independent in ownership [22]. Scholarly interest in the drivers, governance mechanisms, and performance implications of strategic alliances has grown substantially within the field of strategic management.

In exploring the formation motives of strategic alliances, scholars have proposed multiple theoretical frameworks. According to transaction cost theory, strategic alliances serve as a hybrid governance form situated between markets and hierarchies, aiming to mitigate uncertainty and opportunism in transactions [23,24]. Following this, the resource-based and knowledge-based perspectives argue that firms form alliances to complement resources and integrate capabilities, enabling access to rare and hard-to-imitate external assets [25,26]. While the resource-based view highlights the benefits of alliances in accessing complementary resources, it also implies potential over-reliance on partners, which may weaken firms’ independent sustainability capabilities. Social network theory focuses on the embeddedness of social relationships, arguing that the level of trust and the structure of networks play a decisive role in partner selection [22,27]. Additionally, institutional theory also provides an important complement to alliance research [28]. It suggests that under stricter institutional environments, alliances help firms respond to external institutional pressures, thereby gaining legitimacy and social acceptance. Recently, the ecosystem perspective and dynamic capabilities theory [29] have gained growing attention. These frameworks extend the strategic interpretation of alliances by highlighting their role in enabling technology integration and driving systemic innovation.

Challenges such as information asymmetry and lack of trust are often observed in strategic alliances, largely as a result of organizational heterogeneity. To address such challenges, early studies emphasized the importance of contractual governance, aiming to mitigate opportunistic behavior through detailed legal agreements and incentive mechanisms [24]. As cooperative relationships grow increasingly complex, scholars have gradually placed greater emphasis on relational governance. Trust, reciprocity, and social norms are now viewed as essential safeguards for sustaining long-term collaboration [27,30]. In recent years, digital technologies such as artificial intelligence and blockchain have developed rapidly. As a result, technology-enabled governance mechanisms have begun to emerge. These mechanisms improve the flexibility and responsiveness of alliance governance. They also make alliances better suited to fast-changing external environments.

In practice, the role of strategic alliances in promoting innovation and enhancing firms’ market competitiveness has been widely recognized. Through resource sharing and knowledge flows, alliances help firms overcome technological bottlenecks, promote organizational learning, and accelerate capability accumulation [31,32]. In the market domain, alliances can enhance brand synergy and promote market integration. They not only improve the market penetration of new products but also increase brand awareness and the speed of market response [33,34]. Furthermore, strategic alliances are increasingly becoming important platforms for firms to participate in industry governance and standard setting. This significantly enhances their institutional voice within the industrial ecosystem [35].

In summary, strategic alliances are not only a means for firms to overcome resource and capability constraints but also a critical strategic approach for achieving technological synergy, enhancing institutional influence, and achieving long-term economic value. Hence, strategic alliances have become a key pillar in constituting sustainable competitive advantage. However, it remains unclear whether such alliances can indeed play a powerful role in addressing climate risk.

3. Theoretical Mechanisms and Research Hypotheses

With the strengthening of global climate governance, an increasing number of firms are voluntarily choosing to enhance their adaptive and mitigation capacities by participating in strategic alliances. Nevertheless, when embedding alliance decisions within firm-level practices, an individual firm may face heightened climate vulnerability due to its own organizational constraints and deficiencies in alliance mechanisms.

Initially, internal organizational constraints may be one of the key factors undermining a firm’s capacity to address climate risks. Participants in strategic alliances are typically industry leaders, which are often characterized by large scale and complex operations. As a result, they tend to adopt highly bureaucratic organizational structures. Although such organizational structures facilitate the division of responsibilities, they inevitably lead to issues such as long decision-making chains and complex approval procedures. In turn, the firm’s ability to effectively promote and implement alliance policies will be compromised. Therefore, Hoffman [36] points out that there is a certain degree of mismatch between firms’ internal capabilities and the expectations placed on them within alliances. This kind of misalignment may prevent firms from accurately identifying the strategic focus of the alliance, clarifying resource allocation priorities, and ultimately concentrating limited resources on critical areas. At the same time, when firms are unable to consistently fulfill their alliance responsibilities due to capability misalignment, it will lead to a series of counteractive mechanisms, such as trust erosion, resource isolation, and so on. Ultimately, firms’ capacity to address climate risks will be undermined.

Second, corporate culture is a deeply rooted institutional factor that serves as a soft yet non-negligible barrier. Due to the enduring and change-resistant nature of corporate culture, employees often tend to maintain existing cognitive frameworks and operational routines [37]. By nature, alliances often serve as vehicles for institutional transformation and strategic change. Additionally, Piderit [38] considers that resistance to organizational changes among employees often arises from anxiety over uncertainty, possible harm to personal benefits, and limited engagement in the process. Consequently, such cultural inertia is likely to become a critical factor constraining the value realization of alliances. Simultaneously, organizational culture in many firms is typically risk-averse and short-term oriented. In contrast, alliances often aim to address issues that involve high complexity, high uncertainty, and long return cycles. Ultimately, the misalignment of values may act as a barrier to the strategic integration of alliance initiatives, preventing firms from actively developing their environmental resilience and adaptability.

Beyond firm-level factors, deficiencies in alliance mechanisms can also exacerbate the climate vulnerability of participating firms. On the one hand, information asymmetry within the alliance may lead to overly optimistic assessments of alliance progress. Such misjudgments can weaken firms’ awareness of climate risks and reduce their motivation to strengthen adaptive capacity. On the other hand, alliance members often differ significantly in their motivations for participation and in the extent to which they fulfill their commitments. This heterogeneity may undermine the alliance’s capacity for coordinated governance [39]. It may also encourage free-riding behavior. All these factors will eventually lead to the erosion of alliance value. In conclusion, the actual effectiveness of strategic climate alliances is constrained by internal structural barriers within firms and institutional deficiencies in alliance design. Based on this, the following hypothesis is proposed:

H1. 

Holding other factors constant, engaging in strategic alliances may increase firms’ vulnerability for climate adaptation.

According to Upper Echelons Theory, the personal experiences and value orientations of corporate executives play a significant role in shaping their strategic decision-making paths. For example, executives with advanced education or environmental-related experience are more likely to show strategic foresight and a long-term value system [40,41]. Additionally, those raised in severely polluted environments may develop a deeper awareness of ecological responsibility [42]. Moreover, female executives tend to be more proactive in promoting corporate social responsibility and environmental issues, often exhibiting a preference for consensus building in pursuit of long-term sustainability. Accordingly, the cognitive frameworks and value systems of top executives serve as a key starting point for shaping corporate green strategies.

In the context of actual alliance participation, the limited attention resources of corporate executives become increasingly fragmented due to the complexity of multi-party coordination, the expansion of responsibility scopes, and conflicting goals within the alliance members. The attention-based view suggests that what decision-makers focus on determines what actions organizations take [43]. In practice, executive attention tends to concentrate on environmentally symbolic, operationally simple, and highly visible “window-dressing” activities, such as green slogans and promotional campaigns, boilerplate environmental disclosure, and the like. In contrast, firms tend to marginalize efforts to build environmental resilience, as such initiatives require long-term investment of significant capital and involve substantial uncertainties. This attentional shift is mainly characterized by symbolic support for environmental issues, accompanied by reduced prioritization of environmental goals in actual resource distribution and strategy execution. Unfortunately, the tendency to rely on symbolic commitments rather than real investments diminishes the strategic impact of environmental initiatives. The concealment of underinvestment in environmental matters can ultimately undermine a firm’s climate adaptability and long-term resilience. Therefore, without substantive governance and a shared value system, alliances as multi-stakeholder platforms may further weaken firms’ internal capacity to cope with climate risks.

In conclusion, the cognitive characteristics and value orientations of top executives serve as a key basis for shaping corporate green strategy. Yet, when faced with limited cognitive capacity and a complex alliance context, executives are likely to engage in opportunistic behavior. Thus, this not only weakens the effectiveness of environmental governance in practice but also undermines the organization’s long-term adaptability and sustainability in response to climate change. Based on this, the following hypothesis is proposed:

H1a. 

Strategic alliances may fragment executives’ attention to environmental issues.

Moreover, strategic alliances may further weaken firms’ sustainability capacity through the incentive misalignment within organizations. Drawing on Oliver [44], firms may shift green responsibilities to the alliance level as a way to symbolically meet external pressures while reducing internal commitment. As a result, this may lead firms to deprioritize resource commitment to green capacity building and soften the strictness of associated performance metrics. Due to the lack of tangible incentives linked to green objectives, employees show reduced willingness and effort in building sustainability capabilities [45]. In summary, incoherence in organizational reward structures may weaken efforts to build green competencies, thus fostering a reliance on superficial and symbolic sustainability actions. Based on this, the following hypothesis is proposed:

H1b. 

Strategic alliances may undermine their sustainability capacity.

Furthermore, sustainability capability is one of its core competencies for a firm to cope with climate-related uncertainties. This capability is reflected in the firm’s ability to effectively integrate environmental, social, and governance factors into resource allocation, technological innovation, and strategic execution [46]. In addition, the development of sustainability capability allows firms to navigate escalating regulatory scrutiny [44] and meet the surging demand for environmentally responsible consumption [47].

Strategic alliances may fall short of delivering the intended improvements in sustainability capability. Alliance members typically allocate roles according to their strengths within the partnership network. This kind of division of labor can improve operational efficiency in the short run. Nonetheless, Lavie [48] argues that over-reliance on external collaboration in the long run may erode firms’ internal learning processes and reduce their ability to embed green practices across business functions, ultimately narrowing the space for sustainability capability building. Simultaneously, firms tend to institutionalize their green positioning through structured resource deployment, employee training, and assessment frameworks. Yet, such a structured approach may lock firms into path-dependent practices, making it more costly to adapt or revise their green strategies. Thus, the potential risks and costs of reconstructing green capabilities may further reduce firms’ motivation to move beyond their current limiting structures. In the long run, the persistence of such rigid institutional routines may erode adaptability to external environmental changes.

Seemingly, strategic alliances provide firms with a platform for risk-sharing, particularly in responding to external shocks such as supply chain disruptions or market uncertainties. Barratt [49] and Das and Teng [50] claim that such a collaborative mechanism provides a certain buffering effect. In the short term, this mechanism appears to help firms stabilize the environment. Instead, this perceived stability may conceal deeper cognitive distortions in the long term.

Moral hazard theory suggests that individuals tend to reduce their self-protective efforts when external safeguards are in place. Correspondingly, due to reliance on support provided by alliance partners, a firm may neglect independent evaluation and proactive responses to climate risks, resulting in reduced emphasis on investment in adaptive capabilities [51]. As risk awareness diminishes, both the organization’s adaptive capacity and its willingness to act tend to decline.

Crucially, this decline in cognitive vigilance is not limited to top management but can manifest as structural reliance embedded at the organizational level. For instance, due to concerns over knowledge leakage or competitive risks, firms may deliberately avoid sharing high-quality data or engaging in deep collaboration within the alliance, thereby limiting their ability to identify and respond dynamically to climate risks. Meanwhile, resource dependence and path-locking effects within alliances may gradually erode firms’ capabilities, making them less sensitive to external shifts [52,53].

Over time, such dependency from long-term cooperation may significantly undermine a firm’s own capabilities. Instead of independently strengthening their adaptive systems, firms become increasingly reliant on the support offered by their strategic allies. Though effective in boosting short-term performance, this model may lock firms into rigid risk-handling routines over the long term, giving rise to institutional path dependence and undermining organizational agility and innovation [54].

Overall, strategic alliances may generate a false sense of security and undermine both information flow and capability development. As a result, firms may become less sensitive to supply chain risks, ultimately increasing their exposure to climate-related risks. Based on this, the following hypothesis is proposed:

H1c. 

Strategic alliances may reduce firms’ sensitivity to supply chain risks.

4. Research Design

4.1. Sample Selection and Data Sources

This paper uses panel data of A-share listed Chinese firms from 2010 to 2023 as the sample, and then the sample is screened based on the following criteria: (1) Due to the unique characteristics of the financial and insurance industries, financial listed companies are excluded; (2) Companies labeled as ST or ST usually have abnormal financial performance, so they are also excluded; (3) To ensure sample completeness and enhance the reliability of the estimation results, it excludes observations with missing values in key variables; (4) To mitigate the influence of outliers, all continuous variables are winsorized at the 1st and 99th percentiles of their empirical distributions. Unless otherwise noted, all variables are defined as described above. The financial and corporate governance data used in this study are obtained from the Wind and CSMAR databases. As a consequence, the dataset consists of 18,486 observations.

4.2. Definition of Main Variables

4.2.1. Dependent Variable

Climate Risk (ClimateRisk): Referring to Du et al. [55], this study uses the annual reports of A-share listed companies in China from 2010 to 2023 as samples. Then, data sources, including the China Meteorological Disaster Yearbook, the National Meteorological Science Data Center, and the English climate risk vocabulary developed by Li et al. [56], will also be employed. Through text analysis, 76 high-frequency climate risk seed words are identified, which form the basis of the preliminary vocabulary. On the basis of preliminary vocabulary, machine learning techniques are used to improve vocabulary coverage and semantic accuracy. The CBOW (Continuous Bag-of-Words) model is applied to expand the seed words through word embeddings. Related words with the highest similarity are extracted. In the end, 98 climate risk keywords are selected, such as “low-carbon”, “nuclear power”, and so on. Finally, a Climate Risk Index is developed based on the ratio of climate risk word frequency to the total word count in annual reports, serving as a proxy for a firm’s vulnerability to climate change. The higher the index value, the more sensitive the firm is to climate change and the greater its associated vulnerability.

4.2.2. Independent Variable

Strategic alliances status of firms (Alliance): According to Huang et al. [57], this paper constructs the explanatory variable to capture the strategic alliance status of firms based on the alliance announcement disclosed by listed companies. Because not all announcements clearly specify the duration of cooperation, this paper adopts the following treatment for different scenarios: (1) If the announcement discloses a specific duration of the strategic alliance, it is assumed that the alliance will last for the entire agreed period. For example, if a five-year cooperation period is specified in the announcement, the alliance is assumed to remain effective for five consecutive years beginning from the disclosed start time; (2) Alliances that do not specify a cooperation period in the announcement are assumed to last for three years [58]. On the basis of these two principles, it constructs a dummy variable, Alliance. If the company is under the status of the strategic alliances, Alliance equals 1, and 0 otherwise.

4.2.3. Mediating Variable

Executives’ environmental attention (CEOattention): According to Wu and Hua [59], this paper conducts a textual analysis of CSR reports issued by A-share listed companies from 2010 to 2023. Drawing on the Wingo Financial Text Platform and relevant literature, it constructs an environmental keyword dictionary covering themes such as “workplace safety”, “pollution control”, “energy conservation and emission reduction”, “clean energy”, “sustainable development”, and so on. In order to ensure the representativeness of the dictionary, it must comprehensively cover environmental technologies, pollutant types, governance measures, and green development concepts. The indicator is calculated as the ratio of the total frequency of environmental keywords to the total number of words in the CSR report and then scaling it by 100. A higher value of this index indicates greater executive attention to environmental issues.

Corporate sustainability capabilities (Sustainability): The specific calculation formula is presented in

Table 1. Definition of key variables.

Variable Type	Variable Name	Variable Symbol	Variable Description
Dependent Variable	Climate risk	ClimateRisk	Detailed definition is shown in Section 4.2.1.
Independent Variable	Strategic alliances status of firms	Alliance	Detailed definition is shown in Section 4.2.2.
Mediating Variables	Executives’ environmental attention	CEOattention	Detailed definition is shown in Section 4.2.3.
	Corporate sustainability capabilities	Sustainability	Sustainability Capability = (Net Profit Margin × Retention Ratio × (1 + Equity Multiplier))/[(1/Total Asset Turnover) − (Net Profit Margin × (1 + Equity Multiplier))]
	Perception of supply chain risks	RiskPerception	Detailed definition is shown in Section 4.2.3.
Control Variables	Book-to-Market ratio	BM	Book value of Equity/Market value of Equity
	Account receivable ratio	Rec	Accounts receivable/Total assets
	Leverage ratio	Lev	Total liabilities/Total assets
	Tangible asset ratio	TAR	(Total assets-Net intangible asset -net goodwill)/Total assets
	Management expense ratio	Mfee	Management fees/Total assets
	Return on assets	ROA	Net profit/Total assets
	Big 4 auditor status	Big4	Dummy variable equals 1 if their financial reports are audited by the Big Four accounting firms and 0 otherwise
	Herfindahl index	Herfindahl3	Sum of the squared shareholding ratios of the top three shareholders

. Corporate sustainability capability refers to a company’s long-term profitability and competitive strength. Currently, the most widely used approach is Hesketh’s sustainable growth model, which considers sustainable growth as a function of long-term growth and corporate financing behavior. However, this model does not account for differences in the operating environments of companies or issues such as corporate management growth. To address these limitations, the method proposed by Yang et al. [60] is often applied. This method follows a sustainable development framework and uses corporate performance indicators from listed companies to measure sustainability capabilities. The formula involves the comprehensive calculation of several key financial indicators, including net profit margin, retention ratio, equity multiplier, and total asset turnover. It aims to assess a company’s overall ability to utilize resources, achieve profitability, and maintain self-financing cycles, thus reflecting its potential for sustainable development in the face of external challenges.

Perception of supply chain risks (RiskPerception): As firm annual reports rarely provide direct disclosure of risks arising from specific customers or suppliers, we adopt the approach proposed by Chen and Fan [61] to quantify supply chain risk perception. First, we use a Python-based program (V3.10) to collect the annual reports of A-share listed companies from 2010 to 2023 via the iFinD (V3.10). The PDF files are then converted into TXT for subsequent textual analysis. During the text-cleaning process, we remove invalid characters such as images, blank spaces, numbers, and garbled symbols, and segment the text into individual sentences. Subsequently, Chinese word segmentation is performed using Jieba. Based on prior literature and existing studies, we construct a custom dictionary containing keywords such as “risk”, “uncertainty”, “supplier”, “customer”, “procurement”, and so on. The dictionary is designed to cover two major themes: risk-related and supply chain-related terms. Additionally, Word2Vec is employed to semantically extend the initial keyword list by identifying words with high similarity to core terms. The extended terms are cross-validated against a government corpus from the Wingo to ensure contextual relevance and sufficient usage frequency. This is to ensure that the final vocabulary remains both risk-related and supply chain-related. In the final step, it is to identify how often supply chain-related and risk-related keywords appear together within the same paragraph in the annual report. The co-occurrence of supply chain-related and risk-related terms within a short segment is treated as an indication that the firm is addressing supply chain risk topics.

Based on above explanation, the supply chain risk perception index is computed using the following formula:

$${RiskPerception}_{it}=100\times {\displaystyle \frac{1}{N_{i,t}}}\sum _{j=1}^{N_{i,t}}\delta ({\omega }_jϵS)·\delta (\underset{\mathit{kϵR}}{\mathit{min}}\left|j-k\right|\le d)$$

(1)

where $N_{i,t}$ is the total number of words in the annual report of firm $i$ in year $t$; ${\omega }_j$ is the j-th word; $S$ is the set of supply chain-related keywords; $R$ is the set of risk-related keywords; $d$ is the context window size. According to Chen and Fan [61], the window size in this paper is set to 15 words before and after the focal word. Moreover, $\delta (·)$ is an indicator function that equals 1 if the condition inside the parentheses is satisfied, and 0 otherwise; $\underset{\mathit{kϵR}}{\mathit{min}}\left|j-k\right|$ denotes the distance from the j-th word to the nearest risk-related word in the text.

4.2.4. Control Variable

With reference to Yu et al. [62] and Dong et al. [63], the control variables selected for this empirical testing are relevant financial and corporate governance indicators. They are book-to-market ratio (BM), accounts receivable ratio (Rec), leverage ratio (Lev), tangible asset ratio (TAR), management expense ratio (Mfee), return on assets (ROA), Big 4 auditor status (Big4), and the Herfindahl index (Herfindahl3), respectively.

4.3. Model Design

In order to investigate the impact of strategic alliances on the vulnerability of firms in addressing climate change, it set up the economic model as follows:

$${ClimateRisk}_{i,t}={\beta }_0+{\beta }_1{Alliance}_{i,t}+{\beta }_{2}Controls+\sum {Year}_i+\sum {Ind}_i+{\epsilon }_{i,t}$$

(2)

where ${ClimateRisk}_{i,t}$ represents the level of climate-risk exposure for the $i$-th company in year $t$. ${Alliance}_{i,t}$ is the dummy variable, which is to indicate whether the $i$-th company is in an alliance in year $t$. If $i$-th company is in an alliance in year $t$, ${Alliance}_{i,t}$ equals 1. Otherwise, it is 0. To control unobservable factors related to industry characteristics and time effects, this study also controls fixed effects for industry (${Ind}_i$) and year (${Year}_i$). Finally, ${\epsilon }_{i,t}$ is error term and clustered at the firm level.

5. Empirical Results Analysis

5.1. Descriptive Statistics

Descriptive statistics are shown in

Table 2. Descriptive Statistics.

Variables	Mean	SD	Median	Min	Max
ClimateRisk	0.172	0.191	0.111	0.019	1.123
Alliance	0.352	0.478	0.000	0	1
BM	0.349	0.166	0.328	0.035	0.797
Rec	0.122	0.103	0.100	0	0.466
Lev	0.414	0.209	0.404	0.05	0.905
TAR	0.93	0.085	0.959	0.538	1
Mfee	0.088	0.073	0.069	0.008	0.456
Big4	0.06	0.238	0.000	0	1
Herfindahl3	0.156	0.115	0.126	0.012	0.559
ROA	0.035	0.065	0.037	−0.264	0.194

. According to Table 2, the mean of ClimateRisk is 0.172 with a standard deviation of 0.191. It indicates that most firms have exposure to a low to moderate level of climate risk, but the level of exposure varies significantly across companies. Additionally, the median of ClimateRisk is 0.111, which is lower than the mean of ClimateRisk. So, it is a right-skewed distribution of ClimateRisk, suggesting that some firms are exposed to extremely high levels of climate risks. In addition, the mean of the alliance variable is 0.352, with a standard deviation of 0.478. Additionally, Alliance is a dummy variable, so its minimum value is 0 and maximum value is 1. The remaining control variables show reasonable levels of dispersion, reflecting that the sample is reasonably representative.

5.2. The Baseline Regression Result

Table 3. The baseline regression.

	(1)	(2)	(3)
	ClimateRisk	ClimateRisk	ClimateRisk
Alliance	0.030 ***	0.028 ***	0.008 ***
	(5.876)	(5.628)	(3.451)
BM		0.115 ***	0.038 ***
		(5.543)	(3.055)
Rec		−0.066 **	−0.005
		(−1.965)	(−0.176)
Lev		0.181 ***	0.044 **
		(7.765)	(2.522)
TAR		0.015	−0.021
		(0.408)	(−0.735)
Mfee		−0.151 ***	−0.022
		(−3.279)	(−0.764)
Big4		−0.001	0.012
		(−0.043)	(0.912)
Herfindahl3		0.034	−0.005
		(1.022)	(−0.117)
ROA		0.027	0.089 ***
		(0.669)	(4.187)
_cons	0.173 ***	0.060	0.169 ***
	(47.411)	(1.539)	(6.017)
N	18,486	18,486	18,435
r1
r2_a	0.318	0.337	0.846
F	34.528	12.066	4.417

t statistics in parentheses, ** p < 0.05, *** p < 0.01.

shows the result of baseline estimation. Column (1) is the result of estimation without adding any control variables. Based on the baseline model in column (1), column (2) includes control variables and also controls for industry and year fixed effects. In order to capture the firm-level heterogeneity, column (3) further controls firm fixed effects based on column (2). According to Table 3, the core independent variable in all three models is Alliance, which is positively significant at the 1% level. This finding supports H1 that engaging in strategic alliances increases firms’ exposure to climate risk.

5.3. Robustness Check

5.3.1. Instrumental Variable

The problem of endogeneity bias arises due to omitted variables and reverse causality between ClimateRisk and Alliance. In order to address this issue, two-stage least squares (2SLS) will be employed as the instrumental variable approach. Additionally, SubsidiariesNumber is selected as an instrument variable and is the natural logarithm of the total number of subsidiaries controlled by the listed company in that year plus one [64]. Firstly, firms with more subsidiaries often have higher demands for organizational coordination and resource integration. As a result, they are more inclined to join the alliance to improve managerial efficiency and strategic synergy, thereby forming a strong alliance motivation. Secondly, TCFD [2] points out that the climate risks faced by firms are regarded as an exogenous shock, driven by external factors such as physical climate events, regulatory changes, and broader environmental trends. These factors are beyond the control of individual firms. Consequently, the inherently exogenous nature of climate risk indicates that it is primarily shaped by the external environment rather than internal firm characteristics. Accordingly, SubsidiariesNumber satisfies both relevance and exclusion restrictions of instrumental variables.

The result of the first stage is shown in column (1) of

Table 4. Instrumental variable.

	(1)	(2)
	Alliance	ClimateRisk
SubsidiariesNumber	0.0867 ***
	(0.00705)
Alliance		0.346 ***
		(0.0586)
BM	−0.250 ***	0.175 ***
	(0.0446)	(0.0299)
Rec	0.0766	−0.0742 *
	(0.0617)	(0.0381)
Lev	−0.0904 **	0.156 ***
	(0.0441)	(0.0263)
TAR	−0.191 ***	0.113 **
	(0.0623)	(0.0486)
Mfee	0.0485	−0.168 ***
	(0.0998)	(0.0578)
Big4	−0.0417	−0.00311
	(0.0279)	(0.0174)
Herfindahl3	−0.221 ***	0.117 ***
	(0.0574)	(0.0406)
ROA	−0.426 ***	0.133 **
	(0.0914)	(0.0544)
_cons	−0.250 ***
	(0.0446)
Kleibergen-Paap rk LM statistic		131.105
Cragg-Donald Wald F statistic		365.885
Kleibergen-Paap rk Wald F statistic		151.087
N	16,528	16528

t statistics in parentheses, * p < 0.1, ** p < 0.05, *** p < 0.01.

. The coefficient is 0.0867 and statistically significant at the 1% level. This indicates that there is a strong correlation between SubsidiariesNumber and Alliance, thereby statistically verifying relevance restrictions of instrumental variables.

Additionally, column (2) of Table 4 is the result of the second stage. The Kleibergen–Paap rk LM is 131.105 and significant at the 1% level, which indicates that there is no under-identification issue for the instrumental variable. Furthermore, the Cragg–Donald Wald F is 365.885, far exceeding the most stringent Stock–Yogo critical value of 16.38. Therefore, the rejection of the null hypothesis indicates that SubsidiariesNumber is a strong instrumental variable. In summary, the robustness of H1 is empirically supported.

5.3.2. Placebo Test

Figure 1 provides a visual representation of the permutation test results. The x-axis represents the estimated coefficient values. The y-axis on the left represents the p-values, while the y-axis on the right shows the kernel density estimates. Generally, it confirms the robustness of the baseline regression estimates. Specifically, this paper randomly permutes the treatment variable 1000 times, while the dependent variable and control variables are held constant. Then, a distribution plot of the regression coefficients is generated based on the estimated coefficient of the treatment variable from each iteration.

According to Figure 1, the randomly generated coefficients are normally distributed with a mean near zero. It suggests that permutation test results are statistically reliable, as no outliers or systematic bias were detected. In addition, most of the coefficients are not statistically significant at the 10% level, indicating that the null hypothesis cannot be rejected. In other words, strategic alliances will not exacerbate firms’ climate vulnerability. Overall, the result shows a reverse validation for the robustness and reliability of H1.

5.3.3. PSM

In order to enhance the credibility of H1, this paper adopts the PSM method. A radius matching algorithm uses BM, Size, Rec, Lev, TAR, Mfee, AuditFee, Big4, and Herfindahl3 as matching covariates. Additionally, the caliper width is set at 0.05. After passing the covariate balance test for the matched sample, the baseline regression model is re-estimated using the matched data. The regression results are presented in column (1) of

Table 5. Robustness Check.

	PSM	Replace Core Variables
	(1)	(2)	(3)
	ClimateRisk	ClimateRisk	ClimateRisk_new
Alliance	0.029 ***		0.001 ***
	(5.068)		(5.243)
Alliance_new		0.031 ***
		(5.528)
BM	0.126 ***	0.115 ***	0.004 ***
	(5.262)	(5.553)	(4.959)
Rec	−0.076 **	−0.064 *	−0.003 ***
	(−2.228)	(−1.912)	(−2.603)
Lev	0.188 ***	0.180 ***	0.006 ***
	(7.637)	(7.732)	(7.217)
TAR	0.000	0.014	0.006 ***
	(0.005)	(0.392)	(5.029)
Mfee	−0.163 ***	−0.149 ***	−0.008 ***
	(−3.259)	(−3.246)	(−4.116)
Big4	−0.011	−0.001	0.002 *
	(−0.686)	(−0.078)	(1.699)
Herfindahl3	0.032	0.036	0.001
	(0.901)	(1.080)	(0.790)
ROA	0.007	0.029	−0.002
	(0.137)	(0.719)	(−1.424)
Constant	0.072 *	0.057	0.001
	(1.655)	(1.452)	(0.997)
Year	Yes	Yes	Yes
Industry	Yes	Yes	Yes
N	9636	18,486	18,412
R-squared	0.34	0.338	0.245

t statistics in parentheses, * p < 0.1, ** p < 0.05, *** p < 0.01.

.

According to column (1) of Table 5, the coefficient of Alliance is significantly positive at the 1% level, which shows that strategic alliances increase firms’ exposure to climate risks. This result remains consistent with the estimated direction and significance level of the Alliance variable in the baseline regression model, thereby verifying the robustness of H1.

5.3.4. Replace the Independent Variable

In order to verify the robustness of H1, this paper follows Chen et al. [65] in defining a new dummy variable, denoted as Alliance_new. Based on the firm’s strategic announcement, if the strategic alliances started within the past three years, Alliance_new is 1. Otherwise, Alliance_new is 0.

According to column (2) of Table 5, the coefficient of Alliance_new is positive and significant at the 1% level. It suggests that firms experience a significant increase in climate risk exposure during the period of a strategic alliance. Therefore, this result is consistent with the estimated direction and significance of the Alliance in the baseline regression model, further supporting the robustness of H1.

5.3.5. Replace the Dependent Variable

Referring to Li et al. [56] and Yuan [66], this paper uses the density of the keyword related to climate risks in the Management Discussion and Analysis (MD&A) as an alternative measure of ClimateRisk, denoted as ClimateRisk_new. Specifically, it is calculated as the frequency of keyword occurrences divided by the total length of the MD&A text. A higher ratio indicates more emphasis on climate risk in corporate disclosures. So, it also implies that firms may face relatively greater exposure to climate risk.

Column (3) of Table 5 reports the regression results after replacing ClimateRisk with ClimateRisk_new. According to column (3), the coefficient of Alliance is 0.001 and significant at the 1% level. The result is consistent with the baseline regression model in both the sign and the level of statistical significance, further validating the robustness and credibility of H1.

6. Mechanism Test

Drawing on the preceding theoretical framework and empirical findings, this paper further explores the underlying mechanisms through which strategic alliances influence firms’ exposure to climate risk. Accordingly, this paper analyzes the mechanisms through which strategic alliance participation may amplify firms’ exposure to climate risk from three perspectives: the reduction in executives’ environmental attention, the weakening of corporate sustainability capabilities, and the diminished perception of supply chain risks. Corresponding variables are introduced for mechanism testing, and Model (3) is constructed to test H1a, H1b, and H1c.

$${Mechanism}_{i,t}={\alpha }_0+{\alpha }_1{Alliance}_{i,t}+{\alpha }_{2}Controls+\sum {Year}_i+\sum {Ind}_i+{\epsilon }_{i,t}$$

(3)

where ${Mechanism}_{i,t}$ represents variables used for mechanism testing, including proxies for executives’ environmental attention, supply chain risk perception, and sustainability capability.

6.1. Executives’ Environmental Attention

Column (1) of

Table 6. Mechanism test.

	(1)	(2)	(3)
	CEOattention	Sustainability	RiskPerception
Alliance	−0.014 ***	−0.016 **	−0.200 **
	(−6.821)	(−2.216)	(−2.123)
BM	0.011	−0.233 ***	−0.146
	(1.205)	(−8.562)	(−0.302)
Rec	−0.078 ***	0.382 ***	4.186 ***
	(−5.909)	(6.168)	(5.863)
Lev	0.048 ***	0.841 ***	−3.903 ***
	(4.845)	(24.091)	(−8.322)
TAR	0.063 ***	−0.158 ***	2.920 ***
	(4.743)	(−3.741)	(4.052)
Mfee	−0.079 ***	0.283 ***	−8.011 ***
	(−3.987)	(3.992)	(−9.436)
Big4	−0.006	0.008	0.011
	(−1.038)	(0.314)	(0.042)
Herfindahl3	−0.016	−0.129 ***	1.172 *
	(−1.248)	(−3.092)	(1.760)
ROA	−0.070 ***	3.518 ***	−8.131 ***
	(−4.119)	(24.545)	(−9.431)
_cons	0.076 ***	−0.037	7.202 ***
	(5.004)	(−0.781)	(9.054)
N	21,779	22,308	25,989
r1
r2_a	0.313	0.289	0.248
F	18.846	147.596	29.098

t statistics in parentheses, * p < 0.1, ** p < 0.05, *** p < 0.01.

shows the mechanism test result for H1a. According to column (1), the coefficient of Alliance is significantly negative at the 1% level, indicating that strategic alliances reduce executives’ environmental attention. Furthermore, decreased environmental concerns will exacerbate internal vulnerability in responding to climate change.

6.2. Corporate Sustainability Capabilities

Column (2) of Table 6 shows the mechanism test result for H1b. Drawing on column (2), the coefficient of Alliance is significantly negative at the 5% level. The negative coefficient suggests that alliance participation is associated with a downward trend in firms’ sustainability capability. This finding confirms the assumption that strategic alliances may weaken firms’ motivation to invest in sustainability.

6.3. Perception of Supply Chain Risks

Column (3) of Table 6 shows the mechanism test result for H1c. In column (3) of Table 6, the coefficient of alliance is significantly negative at the 5% level, implying that firms’ perception of supply chain risk declines significantly after participating in a strategic alliance. This finding supports the previously proposed “false sense of security” mechanism, whereby the risk-buffering effect provided by strategic alliances may lead managers to subjectively underestimate systemic external risks facing the supply chain. As a result, the “false sense of security” mechanism will weaken the firm’s sensitivity to risk and awareness for prevention.

7. Heterogeneity Analysis

7.1. Heterogeneity Analysis from the Perspective of Carbon Performance

Carbon performance has been widely accepted as a key proxy for assessing climate governance effectiveness, as it reflects both a firm’s greenhouse gas emissions and its climate-related actions [67]. By examining carbon emission intensity, emission reduction effectiveness, and disclosure practices, carbon performance provides a comprehensive reflection of a firm’s actual capacity to implement climate strategies. Accordingly, incorporating carbon performance as a moderating variable in examining the impact of strategic alliances on climate risk enables the identification of heterogeneous effects across firms with varying levels of environmental governance.

The theoretical rationale for selecting carbon performance as a moderating variable is twofold. First, prior studies have found that higher carbon emissions are generally associated with greater financing pressure and higher levels of systemic risk [14]. Thus, carbon performance serves as an indicator of both a firm’s vulnerability to climate change and its resilience in managing climate-related risks. Second, carbon performance serves as a proxy for a firm’s strategic orientation toward environmental management. For example, studies have shown that firms with better carbon performance tend to have higher ESG ratings, more proactive environmental investments, and more systematic sustainability strategies [15].

In the variable construction, this study adopts the method proposed by He et al. [68] and defines the annual carbon performance measure as the reciprocal of the ratio of total carbon emissions to net sales. A higher value of this indicator indicates lower carbon emissions per unit of revenue, reflecting greater carbon efficiency. According to the median of the constructed variable, the sample is divided into a “high carbon performance group” and a “low carbon performance group” for regression analysis.

According to column (1) and (2) of

Table 7. Heterogeneity analysis.

	(1)	(2)	(3)	(4)
	ClimateRisk
	High carbon performance	low carbon performance	High Firm value	Low Firm value
Alliance	0.020 ***	0.039 ***	0.019 ***	0.038 ***
	(3.913)	(5.476)	(3.679)	(5.450)
BM	0.104 ***	0.127 ***	0.084 ***	0.226 ***
	(5.066)	(4.532)	(3.642)	(4.346)
Rec	−0.051 *	−0.088	−0.031	−0.107 **
	(−1.671)	(−1.633)	(−0.875)	(−2.506)
Lev	0.182 ***	0.175 ***	0.133 ***	0.285 ***
	(7.303)	(5.848)	(5.904)	(6.256)
TAR	−0.047	0.124 ***	0.040	−0.014
	(−1.128)	(2.647)	(1.281)	(−0.263)
Mfee	−0.128 ***	−0.149 **	−0.150 ***	−0.165 **
	(−2.830)	(−2.314)	(−4.004)	(−2.082)
Big4	−0.005	0.006	0.010	−0.008
	(−0.369)	(0.273)	(0.631)	(−0.484)
Herfindahl3	0.052	0.015	0.014	0.053
	(1.417)	(0.367)	(0.412)	(1.244)
ROA	0.061	−0.015	0.036	−0.032
	(1.469)	(−0.255)	(0.962)	(−0.467)
_cons	0.113 ***	−0.038	0.057 *	−0.003
	(2.780)	(−0.705)	(1.692)	(−0.046)
N	10577	7906	9240	9246
r1
r2_a	0.364	0.312	0.265	0.377
F	10.749	8.178	7.863	8.932
F-test	−0.019 ***		−0.018 ***

t statistics in parentheses, * p < 0.1, ** p < 0.05, *** p < 0.01.

, the coefficient of Alliance is significantly positive at the 1% level in both groups. Additionally, the F-test indicates that there is statistically significant heterogeneity between negative effects of strategic alliances on climate risks, depending on different carbon performance. This finding suggests that the higher a firm’s value, the greater its ability to manage the environmental uncertainties brought about by strategic alliances in a proactive and prudent manner, thereby effectively controlling the growth of climate risk. Conversely, when firm value is lower, strategic alliances are more likely to be used as a tool for shirking responsibility, which in turn exacerbates the firm’s climate vulnerability.

7.2. Heterogeneity Analysis from the Perspective of Firm Value

Firm-specific characteristics also play a critical role and should not be neglected when assessing the influence of strategic alliances on climate risk. Among these firm-level variables, firm value serves as an important indicator of a company’s market performance and future growth potential. It not only reflects the individual firm’s competence in coordinating resources and absorbing external shocks [69] but also captures the capital market’s overall evaluation of its environmental and climate responsibility [70]. Generally, high-value firms are more active in ESG governance and environmental reputation management. They also exhibit greater adaptability and resilience in response to climate shocks, market volatility, or physical risks.

In contrast, firms with lower market value may be more inclined to leverage strategic alliances as a means of circumventing regulatory obligations, offloading environmental costs, or concealing deficiencies in carbon emission management and environmental compliance. As a result, such behavior may objectively increase their exposure to climate risk.

Drawing from the above, this paper adopts Tobin’s Q as the measure of firm value. Specifically, Tobin’s Q is defined as the ratio of a firm’s market value to its total assets at the end of the period. A higher Tobin’s Q indicates a higher level of market recognition and firm value. Furthermore, the sample is divided into a “high firm value group” and a “low firm value group” based on the sample median of Tobin’s Q. The result is shown in column (3) and (4) of Table 7.

According to column (3) and (4), Alliance is significantly positive at the 1% level in both groups. However, the coefficient is higher for the firms with lower firm value, implying that strategic alliances impose greater climate risk on low-value firms. Additionally, the result of the F-test indicates that the difference between the two is statistically significant.

This empirical result further validates the hypothesis proposed in this study: firms with higher value are better able to respond to the climate risks brought by strategic alliances. In contrast, when firm value is low, strategic alliances tend to serve as tools for superficial compliance or the transfer of environmental responsibilities, which ultimately heightens their exposure to climate risk.

8. Extension Studying

In assessing the impact of strategic alliances on corporate climate risk, differences in alliance types constitute a critical dimension of heterogeneity. This is primarily because different types of alliances fundamentally differ in how they transmit risks, pursue strategic objectives, and are perceived externally.

Equity-based alliances involve equity ties, which foster closer cooperative relationships, higher levels of inter-firm coordination, and more extensive information sharing. However, they also exhibit stronger risk interdependence. Once parts of the alliance members face climate-related risks, those risks can easily spread within the alliance. Additionally, the limited flexibility of equity-based alliances often leads to delayed responses to environmental changes. Moreover, the governance mechanisms of equity-based alliances are more exposed to external regulation and public scrutiny, which makes them more vulnerable to reputation risks.

In contrast, non-equity alliances have a more loosely structured form of cooperation, with clear organizational boundaries and well-defined allocations of risk responsibilities. Their greater adaptability and flexibility enable firms to adjust their cooperative strategies in a timely manner when facing uncertain climate risk. At the same time, such strategic alliances are typically based on contractual arrangements rather than capital ties, allowing firms to explore green technology collaboration and share climate governance resources at a lower cost. This, in turn, enables them to more effectively isolate and manage climate-related risks.

According to the analysis above, this paper divides the companies into two groups. One group includes the companies that participate only in equity-based alliances in that year. The other group involves the companies that participated only in bilateral contractual alliances or those involved in both contractual and equity-based alliances. According to the regression results in columns (1) and (2) of

Table 8. Extension studying: form of alliance.

	(1)	(2)
	Equity-Based Alliance	Non-Equity-Based Alliance
Alliance	0.111 *	−0.049
	(1.905)	(−1.168)
BM	0.023	0.121 **
	(0.197)	(1.994)
Rec	−0.038	−0.157 **
	(−0.253)	(−2.062)
Lev	0.291 ***	0.300 ***
	(3.052)	(5.084)
TAR	0.032	0.096
	(0.253)	(1.166)
Mfee	−0.186	−0.318 **
	(−0.993)	(−2.378)
Big4	0.021	0.042
	(0.344)	(1.129)
Herfindahl3	−0.019	−0.057
	(−0.156)	(−0.705)
ROA	−0.126	0.122
	(−0.721)	(1.521)
_cons	0.015	0.063
	(0.094)	(0.670)
N	390	1111
r1
r2_a	0.446	0.381
F	2.158	4.812
F-test	0.16 ***

t statistics in parentheses, * p < 0.1, ** p < 0.05, *** p < 0.

, Alliance is significantly positive at the 10% level, indicating that there is a positive effect of strategic alliances on climate risk in the equity-based alliance group. In contrast, the coefficient is not statistically significant in the non-equity alliance group.

Further F-tests reveal a statistically significant difference in the regression coefficients between the two types of alliances. This result indicates that the alliance structure chosen by firms plays a critical moderating role in their exposure to climate risk. In equity-based alliances, the close cooperative ties may actually exacerbate firms’ risk exposure. In contrast, the flexible cooperation and adaptive mechanisms of non-equity alliances help firms isolate shocks and enhance their resilience.

In summary, firms engaging in strategic alliances should carefully assess the characteristics of different alliance structures in terms of risk transmission and governance flexibility. Priority should be given to non-equity alliance models, which offer greater adaptability and risk isolation capacity. Such choices can enhance firms’ responsiveness and organizational robustness under conditions of climate uncertainty.

9. Conclusions and Suggestions

Strategic alliances are an important organizational strategy for resource sharing and competitive advantage. Based on panel data of Chinese A-share listed firms from 2010 to 2023, this study shows that strategic alliances can inadvertently increase firms’ exposure to climate risk by diverting managerial attention, weakening sustainability capacity, and creating a false sense of supply chain security. Heterogeneity analysis further indicates that firms with stronger carbon performance and higher firm value are less vulnerable and that non-equity alliances, due to their flexibility, provide greater resilience under climate uncertainty compared to equity-based alliances.

These findings carry clear managerial and policy implications. Firms should include explicit climate-related clauses in alliance contracts to mitigate information asymmetry and reduce risks of free-riding and greenwashing. Environmental performance metrics should be embedded into executive incentives to strengthen managerial focus on sustainability. Moreover, alliance structure matters: while non-equity alliances enhance adaptability, equity-based alliances may amplify risks through deeper interdependence. Firms with weaker carbon performance and lower firm value are particularly exposed, calling for targeted regulatory oversight and improved risk management.

Although this study does not focus on a specific industry, the identified mechanisms—such as attention diversion, weakened sustainability capacity, and underestimated supply chain risks—are broadly applicable across industries. We therefore believe the results provide useful insights for managers and policymakers in diverse sectors. Future work could explore industry-level sub-samples or case studies to generate more tailored recommendations.

This study contributes by (1) constructing a text-based climate risk index, (2) empirically testing three mediating mechanisms, and (3) analyzing heterogeneity in carbon performance, firm value, and alliance structure. These contributions enrich research on climate risk and inter-organizational collaboration while also offering actionable guidance for alliance governance.

Despite these contributions, some limitations remain. The analysis is restricted to Chinese A-share listed firms, which may limit generalizability. In addition, the study relies mainly on quantitative analysis; future research could extend to other contexts and incorporate qualitative methods for deeper insights.