Climate Risk in Supply Chains and Corporate Cash Holdings: Mechanisms and Mitigation Strategies

Abstract

This paper analyzes Chinese A-share listed firms to investigate the influence of climate risks on corporate cash reserves via the supply chain channel. We construct a measure of supplier climate-risk exposure by matching firms’ supplier lists with climate-disaster records. Higher supplier climate risk significantly reduces corporate cash holdings. This suggests that climate-driven cash outflows often exceed firms’ precautionary reserves. Empirical analysis suggests that supplier concentration has a significant moderating effect on outcomes. Firms with more concentrated suppliers tend to detect potential climate risks earlier and establish preventive buffers by augmenting cash reserves, thereby improving financial resilience. We also find that firms’ adoption of AI mitigates the negative cash-flow impacts of climate disasters by improving forecasting and risk-management capabilities.

1. Introduction

Uncovering the impact of supply chain climate risk on firms’ cash holdings helps clarify the mechanisms shaping corporate risk management and financial decisions in the era of global warming. With the increasing frequency of extreme weather events, climate risk faced by suppliers has become a critical external factor threatening operational stability (Pankratz & Schiller, 2024) [1]. In this context, how firms enhance their resilience through cash reserves is an important question. On the one hand, firms may increase cash reserves for precautionary purposes to address emergency purchases, production adjustments, logistical disruptions, and customer defaults caused by supply chain shocks (Opler et al., 1999) [2]. On the other hand, extreme weather conditions can accelerate cash outflows. To enhance supply chain resilience, firms need to invest cash to build distributed production capacity, automated inventory systems, and disaster prevention mechanisms (Wagner & Bode, 2008) [3], creating the reality of the ‘pay to play’ dilemma. In addition, if they cannot obtain timely financing support for high-intensity expenditures such as liquidated damages and emergency purchases, the level of cash holdings will further decline (Pankratz, Bauer, & Derwall, 2023) [4].

Based on the information of the top five suppliers disclosed in the annual reports of Chinese A-share listed companies, this paper constructs a firm-level measure of suppliers’ climate risk exposure by matching suppliers’ locations with local climate-disaster records. We then empirically examine how climate risk affects firms’ cash holdings through supply-chain transmission. The results show that supplier climate risk significantly reduces firms’ cash holdings, indicating that climate-related cash outflows often exceed firms’ precautionary reserves. Further analyses reveal that this negative effect is less pronounced in firms with higher supplier concentration. A more concentrated supply-chain structure enables enterprises to have more accurate expectations of climate risks, thereby reducing the negative impact of shocks on cash holdings. Meanwhile, AI adoption significantly mitigates the impact of climate risk by enhancing firms’ predictive capacity and financial resilience.

From a supply chain perspective, this study investigates how climate risks external to firms influence their financial decisions, and makes three main contributions. First, this research expands the literature on the determinants of corporate cash holdings by incorporating supply-chain climate risk into the framework of corporate financial decision-making. It provides new firm-level evidence on the economic consequences of climate change. Compared with existing studies focusing on internal governance (Harford et al., 2008) [5], financing constraints (Denis & Sibilkov, 2010) [6], and macro volatility (Anand et al., 2018; Orlova, 2020) [7,8], this paper incorporates external environmental factors and demonstrates the significant impact of climate risk on corporate liquidity management. Second, we develop a quantifiable indicator of “supplier climate risk exposure” by linking firms’ supplier information with climate disaster data. This approach provides a new approach to corporate climate risk assessment. Compared with existing research that primarily focuses on measuring climate risk at the macro or industry level and thus has few to capture firm-level exposures along the supply chain. Our study addresses that limitation and offers a more precise method for evaluating corporate climate risk (Ciccarelli & Marotta, 2024) [9]. Third, it reveals the mitigating effect of artificial intelligence. Prior research has largely focused on the direct economic costs of climate risk, paying limited attention to the role of emerging technologies in enhancing firms’ adaptive capacity (Javadi & Masum, 2021; Kabir, Rahman, Rahman, & Anwar, 2021) [10,11]. By highlighting AI adoption, this study shows that AI can strengthen enterprises’ risk perception and response ability, thereby reducing their financial vulnerability under the impact of climate risk.

2. Literature Review and Theoretical Hypotheses

2.1. Literature Review

2.1.1. Climate Risk and Corporate Financial Decision-Making

Climate risk has emerged as a significant factor influencing corporate behavior, exerting profound impacts on firms’ financial stability, investment decisions, and risk management. Existing research indicates that climate-related hazards, such as extreme weather, flooding, and abnormal temperatures, disrupt production processes, damage assets, and elevate operational costs (Cai & Lontzek, 2019; Ginglinger & Moreau, 2023) [12,13]. The heightened uncertainty stemming from climate events also elevates financing costs and cash flow volatility, prompting greater caution in formulating investment and financing policies (Huang, Kerstein, & Wang, 2022; Krueger, Sautner, & Starks, 2020) [14,15]. At the financial policy level, existing literature indicates that climate risk alters corporate cash holding decisions by influencing risk exposure and capital availability (Javadi, Masum, Aram, & Rao, 2023) [16]. Firms located in regions prone to frequent climate shocks or facing heightened environmental regulatory pressure often increase precautionary cash reserves to bolster their resilience against potential operational disruptions and asset losses (Addoum, Ng, & Ortiz-Bobea, 2020) [17]. This behavior reflects firms’ risk-defensive motivation in the face of external uncertainty, namely enhancing resilience to climate shocks by holding greater liquidity resources (Lins, Servaes, & Tamayo, 2017) [18]. However, contrary perspectives have also emerged. Climate-related expenditures, such as green transition costs, emissions reduction investments, or post-disaster restoration expenses, may substantially deplete corporate liquidity, thereby reducing cash holdings (Almaazmi, Al-Shboul, & Barood, 2025) [19]. Furthermore, confronted with long-term climate policy pressures and carbon emission constraints, some firms prefer to allocate capital towards energy structure optimization and environmental governance rather than maintaining substantial cash reserves (Flammer, Hong, & Minor, 2019) [20]. Consequently, the impact of climate risk on corporate cash holdings is not unidirectional but depends on variations in corporate response strategies, financing constraints, and the external policy environment.

2.1.2. Supply Chain Climate Risk and Corporate Resilience

Supply chains serve as a significant conduit for climate risk transmission (Ghadge, Wurtmann, & Seuring, 2020) [21]. When critical suppliers or logistics networks experience disruptions due to climate events, enterprises face production stagnation, rising costs, and delivery delays (Kim, Chen, & Linderman, 2015) [22]. Research indicates that enterprises with geographically concentrated supply chains or high climate exposure are more vulnerable to the adverse impacts of climate shocks (Patatoukas, 2012) [23]. In such circumstances, enhancing supply chain resilience through diversification, redundancy, or long-term stable partnerships is considered a key means of mitigating financial distress stemming from climate risks (Yu, Wu, & Tan, 2025) [24]. Concurrently, supplier concentration plays a significant role in the transmission of external shocks. On the one hand, high supplier concentration implies greater reliance on a small number of critical suppliers; when these suppliers are affected by climate disasters, the associated risks for the firm increase accordingly (Carvalho, 2014) [25]. Conversely, long-term stable supply relationships foster enhanced collaboration and trust between parties, facilitating information sharing that enables earlier identification of potential risks and implementation of effective countermeasures (Freudenreich, Lüdeke-Freund, & Schaltegger, 2020; Tang & Tomlin, 2008) [26,27]. Consequently, supplier concentration may simultaneously confer both risk exposure and protective effects.

2.1.3. Artificial Intelligence, Risk Management and Financial Decision-Making

The application of artificial intelligence is reshaping corporate behavior in risk forecasting, supply chain monitoring, and financial planning. AI technology enables enterprises to process vast quantities of environmental, meteorological, and operational data in real time, thereby enhancing risk early warning and resource allocation efficiency (Baryannis, Dani, & Antoniou, 2019; Farboodi, Mihet, Philippon, & Veldkamp, 2019) [28,29]. Within climate risk scenarios, AI assists enterprises in proactively identifying weather-related disruptions, optimizing logistics, and reserving cash reserves for sudden risks (Duan, Edwards, & Dwivedi, 2019) [30]. In risk management and financial decision-making practices, AI enhances risk identification and forecasting capabilities through data-driven approaches, thereby strengthening corporate decision efficiency and resilience (Brynjolfsson & McElheran, 2016) [31]. Within risk management, AI technologies can integrate multi-source data from meteorological, geographical, logistical, and supply chain domains, thereby heightening corporate sensitivity and responsiveness to external shocks (Agrawal, Gans, & Goldfarb, 2019) [32]. Research indicates that AI applications not only refine the accuracy of risk assessment models but also optimize supply chain scheduling and inventory management through automated algorithms, thereby reducing operational risks and disruption losses (Ghadge et al., 2020) [21]. Within financial decision-making, AI applications bolster enterprises’ information processing capabilities in financing, investment, and cash flow management (Farboodi et al., 2019) [29]. Through deep learning and natural language processing, AI can conduct comprehensive analyses of financial statements, news texts, and market data, aiding businesses in more accurately assessing funding requirements and liquidity risks (Boh, Constantinides, Padmanabhan, & Viswanathan, 2023) [33].

Despite growing research on climate risk and corporate liquidity, the relationship between climate risk, supply chain exposure, and corporate cash holdings remains under-explored. Most studies focus solely on the direct impacts of climate risk on firms or industries, overlooking indirect transmission mechanisms through supply chains. Furthermore, while supplier concentration and AI adoption are widely recognized in strategic management, their moderating roles within this relationship have not been sufficiently empirically tested.

2.2. Theoretical Analysis and Research Hypotheses

Extreme weather events, abnormal temperatures and frequent natural disasters resulting from climate change expose businesses to operational risks such as asset damage, supply chain disruptions and production delays (Krueger, Sautner, & Starks, 2020) [15]. Climate risks often impact not only a firm’s own operations but also propagate through upstream suppliers’ production and logistics chains, thereby amplifying cash flow uncertainty (Bolton & Kacperczyk, 2021; Singh, 2020) [34,35]. Theoretically, firms hold cash for precautionary and transactional motives (Keynes, 1937) [36]. When external uncertainty rises, firms typically increase cash reserves to withstand potential shocks (Bates, Kahle, & Stulz, 2009; Han & Qiu, 2007) [37,38]. However, under climate risk scenarios, firms simultaneously face substantial capital outflows for post-disaster recovery, production restart, and green transformation investments (Krüger, 2015; Lins, Servaes, & Tufano, 2010) [39,40]. Consequently, climate risks may simultaneously motivate firms to increase precautionary cash holdings while actually reducing available cash due to rising costs and operational pressures. The combined effect of these dual mechanisms means that changes in corporate cash reserves depend on the intensity and persistence of climate risk shocks, as well as the firm’s own risk tolerance. In the Chinese context, when climate disasters occur in regions where upstream suppliers are located, firms must not only bear the costs of sourcing alternative suppliers but also contend with cash flow constraints arising from production delays (Wu, Deng, & Gao, 2025) [41]. Consequently, from a supplier perspective, the transmission effects of climate risk are more likely to lead to a reduction in corporate cash holdings, thereby proposing H1.

H1. 

Supplier climate risks can significantly reduce a company’s cash holdings.

Traditional wisdom suggests that high supplier concentration increases a firm’s supply chain risk, as overdependence on a few suppliers may lead to vulnerability to external shocks (Wagner & Bode, 2008) [3]. However, from a “relationship-specific investment” perspective, high supplier concentration not only implies increased dependence, but also represents a shift from market-based transactions to relationship-based governance (Dyer & Singh, 1998; Williamson, 2008) [42,43]. When a firm’s major inputs are concentrated in a small number of core suppliers, the relationship evolves into a strategic partnership, accompanied by substantial relationship-specific investments, creating a deep binding of interests, which provides an intrinsic stabilizing mechanism to cope with climate shocks, and strengthens the firm’s financial resilience and ability to adjust dynamically (Ivanov, 2022; Teece, 2007) [44,45]. First, high supplier concentration strengthens coordination and information sharing between firms and suppliers. With the help of shared information systems, firms can quickly access disaster and recovery information in the event of a shock, jointly formulate contingency plans, and improve information processing and resilience, thus realizing higher liquidity management efficiency at the financial level, avoiding blind, high-cost substitution purchasing, and reducing contingency expenditures (Brandon-Jones, Squire, Autry, & Petersen, 2014; Christopher & Peck, 2004) [46,47]. Second, a stable partnership creates a “quasi-vertical integration” effect, giving both parties an incentive to help each other in the event of a shock. Downstream firms can assist suppliers in resuming production through prepayment or technical support, while suppliers prioritize orders from core customers. This relational mutual aid mechanism reduces firms’ cash flow volatility during crises, and realizes the dynamic repair of timely resource reallocation (Knemeyer, Zinn, & Eroglu, 2009; Wieland & Wallenburg, 2013) [48,49]. In addition, high concentration gives firms greater bargaining power to negotiate additional costs in a crisis, externalizing and smoothing financial stress through risk-sharing mechanisms, and increasing their financial resilience to external shocks (Bhamra, Dani, & Burnard, 2011; Sheffi & Rice Jr, 2005) [50,51]. Finally, firms are more inclined to invest in climate resilience jointly with core suppliers, such as reinforcing plants, building equipment to aid capacity, or sharing inventory. These types of ex-ante planned investment expenditures are predictable, less costly, and more conducive to maintaining cash flow stability than ex-post temporary replacement expenditures, realizing the strategic choice of firms to invest in exchange for long-term financial stability (Ambulkar, Blackhurst, & Grawe, 2015) [52]. In summary, high supplier concentration not only helps firms to maintain operational stability at lower costs during climate shocks by strengthening relationship governance, promoting information sharing, forming mutual aid mechanisms and joint investment, but also enhances their dynamic capabilities in resource allocation, risk mitigation and cash management, thereby strengthening their overall financial resilience, thereby proposing H2.

H2. 

Enterprises with a high degree of supplier concentration experience less impact on their cash holdings from supply chain climate risks.

With the rapid development of Artificial Intelligence (AI) technology, enterprise risk management models are undergoing profound changes. By integrating big data, machine learning, and predictive modeling, AI not only improves enterprises’ ability to predict supply chain climate risks but also serves as a strategic capability that significantly enhances their financial resilience and dynamic adjustment ability, thus mitigating the impact of supply chain climate shocks on their cash holdings (Baryannis, Dani, & Antoniou, 2019) [28]. First, AI enables forecasting and early warning to transform unexpected shocks into predictable events. Businesses can use AI to analyze historical climate data, real-time weather forecasts, and satellite imagery to make highly accurate predictions of extreme weather events at supplier locations (Sarkis, 2020) [53]. Identifying risks in advance gives companies time to prepare by making planned adjustments to purchasing and safety stocks, thus avoiding “panic purchasing” and high emergency logistics costs in the event of a shock, reducing cash outlays and easing liquidity pressures (Ivanov & Dolgui, 2020) [54]. Second, AI can enhance supply chain optimization and resilience. With tools such as supply chain control towers and digital twins, companies can simulate and optimize supply chain networks in real time (Alazab, Alhyari, Awajan, & Abdallah, 2021) [55]. When suppliers are identified as high risk, AI can quickly assess alternatives, optimize logistics routes, and dynamically adjust inventory to maintain service levels with minimal cash burn. At the same time, AI can reduce cash flow volatility by intelligently rescheduling production plans, reducing production downtime and liquidated damages due to raw material shortages (Christopher & Peck, 2004) [47]. Third, AI can facilitate automated response and decision support, transforming chaotic decisions into precise actions. By integrating with ERP, SRM, and other systems or presetting automated response rules, organizations can quickly execute data-driven decisions, shorten response times, and reduce human error (Chen, Chiang, & Storey, 2012) [56]. Orders can be automatically released to alternate suppliers in specific scenarios to quickly capture the optimal response window, thereby reducing additional cash outlays due to delayed or misguided decision-making. In summary, AI technology can mitigate the negative impact of supply chain climate risk on firms’ cash holdings through forecasting and early warning, supply chain optimization, and automated decision making, which strengthens firms’ financial resilience and dynamic adjustment capabilities. Based on this, we propose Hypothesis 3:

H3. 

The application of enterprise artificial intelligence can effectively mitigate the adverse impact of supply chain climate risks on corporate cash holdings.

Figure 1 depicts the research framework underpinning this study:

3. Data and Methodology

3.1. Sample

This paper focuses on China’s A-share listed companies from 2010 to 2022. The main data are drawn from the CSMAR and Wind databases. Using the top five suppliers and their disclosed addresses, we identify the supply regions of each firm at the prefecture-level city or county level. We exclude financial and real estate firms, ST/*ST firms, and observations with missing values for key variables. Continuous variables are winsorized at the upper and lower 1% deciles to reduce the influence of extreme values. After these procedures, the final sample comprises 14,560 quarterly firm observations, which can be found in Supplementary Materials.

3.2. Variables

Supply chain climate risk exposure (CR) is the main independent variable of this study. For each firm, if a climate disaster occurs in a supplier region, the proportion of purchases from that region is considered exposed to supply disruption risk. CR is defined as the weighted average of the ratio of quarterly climate disaster losses to local GDP across the firm’s supply regions, with weights corresponding to the proportion of purchases from each region relative to total purchases, reflecting the firm’s dependence on each supplier region.

Corporate cash holding (Cash) is the dependent variable and is defined as (money funds + trading financial assets)/(total assets − money funds − trading financial assets), following Di et al. (2020) [57]. To eliminate industry- and quarter-specific effects, Cash is demeaned along both dimensions to construct a relative cash holding measure, which better captures firm-level cash management behavior.

Control variables include enterprise size (Size), leverage ratio (Lev), earnings management (EM), short-term solvency (DLCR), return on assets (ROA), net profit margin (NetProfit), quick ratio (Quick), inventory ratio (INV), management shareholding (Mshare), institutional shareholding (FinInst), and chairman shareholding ratio (ChairHoldR), controlling for heterogeneity in financial structure and corporate governance.

Two additional variables are also constructed. Supplier Concentration is measured as the ratio of purchases from the top five suppliers to operating revenues (Barrot & Sauvagnat, 2016) [58]. AI application is a dummy variable identifying AI-related disclosures in the Management’s Discussion and Analysis (MD&A) sections of annual reports using the BERT model. AI takes the value of 1 if AI-related terms appear at least five times and 0 otherwise (see Appendix A), serving as a proxy for firms’ digital capabilities.

3.3. Model Specification

To examine the effect of firm’s exposure to climate change of their supply chain supplier on their cash holdings, we use a multivariate panel model as follows:

$$CashHoldin{g}_{i,t}={\beta }_0+{\beta }_{1}C{R}_{i,t}+{\beta }_2{X}_{i,t}+Fir{m}_i+Yea{r}_t+Industr{y}_i+{ϵ}_{i,t}$$

(1)

where $CashHoldin{g}_{i,t}$ denotes corporate cash holdings, $C{R}_{i,t}$ refers to firm’s exposure to climate change of their supply chain supplier for firm i in time t. $X_{i,t}$ is a vector of control variables. $Fir{m}_i$, $Yea{r}_t$, and $Industr{y}_i$ are the firm, year, and industry fixed effects, respectively. ${ϵ}_{i,t}$ is the error term.

4. Empirical Results

4.1. Descriptive Statistics

Table 1. Descriptive statistics.

	Obs	Mean	SD	Min	Max
CashHolding	14,609	0.0028	0.2117	−0.0898	14.3372
CR	14,609	0.0002	0.0010	0.0000	0.0403
Size	14,609	21.9673	1.2611	15.5773	28.1092
Lev	14,609	0.4434	0.3372	0.0080	10.0822
EM	14,609	2.1708	3.5889	−1.5 × 102	90.5706
DLCR	14,609	0.1494	0.1906	−1.1752	3.0305
ROA	14,609	0.0367	0.1471	−1.5613	7.2493
NetProfit	14,609	0.0541	0.7033	−11.4355	34.2699
Quick	14,609	2.1557	4.2091	0.0524	90.5148
INV	14,609	0.1413	0.1405	0.0000	0.9150
Mshare	14,609	12.3961	19.3260	0.0000	89.1771
FinInst	14,609	0.0711	0.2569	0.0000	1.0000
ChairHoldR	14,609	7.8504	13.7875	0.0000	67.3200

presents the descriptive statistics for key variables. The sample comprises 14,609 quarterly observations across enterprises. The mean value for corporate cash holdings (CashHolding) stands at 0.0028 with a standard deviation of 0.2117, indicating substantial variation in cash reserve levels across firms. The climate risk indicator (CR) exhibits a low mean (0.0002) but a maximum value of 0.0403, suggesting substantial climate exposure risks for some firms. Firm size (Size) has a mean of 21.97 and a standard deviation of 1.26, indicating a relatively concentrated distribution of firm sizes within the sample. The leverage ratio (Lev) has a mean of 0.44, reflecting moderate overall leverage levels across firms. The earnings management indicator (EM) and quick ratio (Quick) exhibit substantial standard deviations, reflecting significant disparities in financial flexibility and earnings manipulation across firms. Furthermore, the mean for supplier climate risk variability (DLCR) stands at 0.1494, indicating pronounced differences in climate risk exposure at the supply chain level. Overall, the distribution of variables aligns with expectations, laying the groundwork for subsequent regression analysis.

4.2. Climate Change and Corporate Cash Holdings

The Baseline regression results are presented in

Table 2. Baseline results.

	(1)	(2)	(3)
	CashHolding	CashHolding	CashHolding
CR	−0.7354 ***	−0.7426 ***	−0.7411 ***
	(0.2723)	(0.2689)	(0.2662)
_cons	0.0029 ***	−0.0696	−0.0714
	(0.0000)	(0.0450)	(0.0453)
Controls	No	Yes	Yes
Firm FE	Yes	Yes	Yes
Year FE	Yes	Yes	Yes
Industry FE	No	No	Yes
N	14,560	14,560	14,559
R2	0.3288	0.3291	0.3291

Note: The report includes robust standard errors that are clustered at the firm level, indicated in brackets. *** p < 0.01; ** p < 0.05; * p < 0.1.

. Column (1) shows a significant negative relationship between supplier climate risk and corporate cash holdings at the 1% level. This result remains robust after adding control variables (Column 2) and industry fixed effects (Column 3). The findings suggest that firms facing higher supply chain climate risk tend to hold less cash. This may be because climate risk disrupts upstream supply stability, raising input costs and operating expenses, which weakens internal capital accumulation. Additionally, to manage supply chain uncertainty, firms may prefer external financing methods, such as prepayments or trade credit over precautionary cash reserves. Consequently, the increase in cash needed for risk response is outweighed by the rise in cash outflows, leading to reduced overall cash holdings. Hypothesis 1 has been proven correct.

4.3. Robustness Check

To test the robustness of the Baseline results, this paper conducts several robust tests. To mitigate the potential endogeneity bias, this paper uses the instrumental variable method for testing. The selected instrumental variable is the difference between the climate risk of an enterprise’s supply chain and the mean value of other enterprises in the same city and industry. Theoretically, the instrumental variable satisfies the correlation and exogeneity conditions. First, firms in the same geographic region and industry face similar business environments and have a common trend in their supply chain risk, so the correlation holds; second, the risk mean of other firms in the same group will not directly affect the firm’s cash holding decision, which is in line with the exclusion constraint. The two-stage least squares regression results in

Table 3. Robustness check: Instrumental variable method.

	(1)	(2)
	First-stage	Second-stage
	CR	CashHolding
CR		−0.7530 ***
		(0.2713)
IV	1.0005 ***
	(0.0008)
Controls	Yes	Yes
Firm FE	Yes	Yes
Year FE	Yes	Yes
Industry FE	Yes	Yes
Kleibergen-Paap rk LM		33.3384
F		1,579,697.5452
N	14,559	14,559

Note: The report includes robust standard errors that are clustered at the firm level, indicated in brackets. *** p < 0.01; ** p < 0.05; * p < 0.1.

show that the F-statistic of the instrumental variable validity test is greater than 10 and significant at the 1% level, rejecting the original hypothesis of weak instrumental variables. After controlling for endogeneity, the negative effect of supply chain climate risk on firms’ cash holdings remains significant, indicating that the conclusions of the benchmark regression are robust. And in

Table 4. Robustness check: Other robustness tests.

	(1)	(2)	(3)	(4)
	CashHolding	CashHolding	CashHolding	CashHolding
CR	−7.42 × 103 ***		−0.8027 ***	−0.5711 ***
	(2.7 × 103)		(0.2816)	(0.2183)
L.CR		−0.6533 *
		(0.3410)
HHI				−0.0254 **
				(0.0121)
SA				−0.0133 **
				(0.0054)
Occupy				0.0049
				(0.0045)
INST				−0.0000
				(0.0000)
_cons	−713.0008	−0.0883	−0.0922	−0.0972 ***
	(451.7518)	(0.0639)	(0.0582)	(0.0324)
Controls	Yes	Yes	Yes	Yes
Firm FE	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes	Yes
N	14,558	10,432	11,556	13,043
R2	0.3291	0.4397	0.3291	0.2504

Note: The report includes robust standard errors that are clustered at the firm level, indicated in brackets. *** p < 0.01; ** p < 0.05; * p < 0.1.

Column (1), we replace the cash holding measure from (money funds + trading financial assets)/(total assets − money funds − trading financial assets) to (money funds + trading financial assets)/(total assets − cash and cash equivalents), which better isolates liquid assets from total assets. The negative impact of supplier climate risk remains significant at the 1% level, confirming that the results are not sensitive to the choice of measurement. In Column (2), to address potential reverse causality, we lag the key explanatory variable supplier climate risk by one period. The lagged variable still shows a significant negative effect on cash holdings, reinforcing the causal interpretation. In Column (3), we exclude samples from the COVID-19 period (2020 onwards) to account for external shocks that may alter firms’ cash behavior. The results remain robust, suggesting that the findings are not driven by pandemic-related anomalies. In Column (4), to rule out the influence of other factors on corporate cash holdings, we further include additional control variables to mitigate the impact of omitted variables. Building on the baseline regression, we add four controls: market competition (HHI), financing constraints (SA), shareholder fund occupancy (Occupy), and institutional investor ownership (INST). The empirical results show that even after accounting for these additional determinants of cash holdings, supply chain climate risk still exerts a negative impact on corporate cash holdings.

4.4. Mechanism Analysis-Based on the Predictive Capabilities of Enterprises

To further examine the relationship between supplier climate risk and corporate cash holdings, this paper incorporates an interaction term between supplier concentration and climate risk to test the moderating effect. Supplier concentration is measured by the ratio of purchases from the top five suppliers to total revenue; higher values indicate stronger dependence on a few suppliers and a more centralized supply chain, implying higher single-point failure risk and lower flexibility in substitution. The regression results in

Table 5. Mechanism analysis-Based on the predictive capabilities of enterprises.

	(1)	(2)	(3)
	CashHolding	CashHolding	CashHolding
Supplier concentration * CR	2.5761 *	2.4997 *	2.1723 **
	(1.3811)	(1.3624)	(1.0671)
CR	−1.9664 **	−1.9330 **	−1.7809 **
	(0.8297)	(0.8301)	(0.6969)
Supplier concentration	0.0073	0.0094	0.0158
	(0.0084)	(0.0107)	(0.0168)
_cons	0.0009	−0.0835	−0.0865
	(0.0023)	(0.0576)	(0.0594)
Controls	No	Yes	Yes
Firm FE	Yes	Yes	Yes
Year FE	Yes	Yes	Yes
Industry FE	No	No	Yes
N	14,508	14,508	14,507
R2	0.3288	0.3291	0.3291

Note: The report includes robust standard errors that are clustered at the firm level, indicated in brackets. *** p < 0.01; ** p < 0.05; * p < 0.1.

show that the coefficient of the interaction term is significantly positive, indicating that higher supplier concentration can significantly mitigate the negative impact of climate risk on corporate cash holdings. This finding suggests that firms with highly concentrated suppliers face higher disruption risks under climate shocks and are better able to anticipate supply chain vulnerabilities due to the stable nature of supplier relationships. Consequently, these firms are more likely to increase cash reserves preemptively to enhance financial resilience. The elevated cash holdings thus reflect not just a reactive measure but a precautionary buffer in anticipation of disruptions. Conversely, firms with more diversified suppliers are less dependent on individual suppliers, but they also have weaker predictive ability for supplier-specific climate risks. As a result, they may fail to recognize early warning signals and are less likely to adjust cash reserves in time, leading to lower cash holdings under climate risk. Overall, the results highlight the importance of supply chain structure in shaping firms’ financial responses to climate uncertainty, with supplier concentration playing a critical role in determining firms’ ability to implement precautionary liquidity strategies. Therefore, the empirical results confirm the validity of Hypothesis 2.

4.5. Mitigation Effect—Analysis Based on Artificial Intelligence Applications

Based on the above analysis, this paper finds that cash reserves driven by risk expectations, especially those generated through forecasts of suppliers’ climate risks, are an effective mechanism for enterprises to cope with climate-related supply chain disruptions. Compared with firms with more diversified supplier networks, those with higher supplier concentration can more easily anticipate climate-induced supply chain shocks, accordingly build up cash reserves in advance to maintain financial flexibility. However, the growing adoption of AI technologies has significantly improved firms’ risk prediction capabilities, enabling them to make anticipatory decisions and engage in proactive financial management.

To explore this mechanism further, this study investigates whether AI application can mitigate the negative impact of supply chain climate risks on firms’ cash holdings. Specifically, three dummy variables are constructed to measure AI usage: (1) General AI application dummy, indicating whether a firm reports the use of any AI-related technology; (2) General-purpose AI dummy, which captures firms applying AI in broad operational areas like data forecasting or process optimization; and (3) Key AI technology dummy, reflecting firms adopting advanced or disruptive AI technologies, such as deep learning and intelligent manufacturing.

The regression results in

Table 6. Mitigation Effect—Analysis Based on Artificial Intelligence Applications.

	(1)	(2)	(3)	(4)	(5)	(6)
	CashHolding (General AI)	CashHolding (NO General AI)	CashHolding (General-purpose AI)	CashHolding (No General-purpose AI)	CashHolding (Key AI)	CashHolding (No Key AI)
CR	−0.4302	−0.5847 **	−0.0028 ***	−0.7450 ***	0.0102	−0.6537 ***
	(0.3855)	(0.2604)	(0.0009)	(0.2679)	(0.0072)	(0.2487)
_cons	−0.1756 ***	−0.0434	−0.0001	−0.0723	−0.0004	−0.0694
	(0.0542)	(0.0477)	(0.0007)	(0.0464)	(0.0003)	(0.0446)
Controls	Yes	Yes	Yes	Yes	Yes	Yes
Firm FE	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes	Yes	Yes	Yes
N	2455	12,038	412	14,083	520	13,980
R2	0.7961	0.3292	0.9450	0.3291	1.0000	0.3293

Note: The report includes robust standard errors that are clustered at the firm level, indicated in brackets. *** p < 0.01; ** p < 0.05; * p < 0.1.

show that supplier climate risk reduces cash holdings only in firms without AI adoption. Further analysis indicates that this mitigating effect of AI is driven by firms using Key AI, while those adopting General-purpose AI remain significantly affected. This is because Key AI is more effective in collecting meteorological data, supplier operational information, and logistics data, enabling firms to identify potential supply chain climate risks earlier and adjust cash reserves in a timely manner to cope with unexpected shocks. In contrast, General-purpose AI has weaker capabilities in data collection and processing, making it less effective in providing timely feedback on climate risks and assisting firms in adjusting their cash reserves. Therefore, the results provide empirical support for Hypothesis 3, confirming its validity.

5. Heterogeneity Analyses

In addition, this paper conducts heterogeneity analyses on the impact of supply chain climate risk on corporate cash holdings, focusing on firm ownership, financing constraints, bank–firm relationships, and the level of regional financial development.

In terms of the nature of firm ownership, the empirical results in

Table 7. Heterogeneity analyses—Based on firm ownership and financing constraints.

	(1)	(2)	(3)	(4)
	CashHolding (SOEs)	CashHolding (Non-SOEs)	CashHolding (High financing constraints)	CashHolding (Low financing constraints)
CR	−0.2423	−1.0305 **	−1.0920 **	−0.0084
	(0.1783)	(0.5017)	(0.4732)	(0.0137)
_cons	−0.0166	−0.1225	−0.2201	−0.0082
	(0.0146)	(0.0778)	(0.1799)	(0.0082)
Controls	Yes	Yes	Yes	Yes
Firm FE	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes	Yes
N	5547	8777	7503	7041
R2	0.5615	0.3301	0.3306	0.7878

Note: The report includes robust standard errors that are clustered at the firm level, indicated in brackets. *** p < 0.01; ** p < 0.05; * p < 0.1.

show that supply chain climate risk has a significant negative effect on cash holdings of non-state-owned firms, while the effect on state-owned firms is not significant. This reflects the fundamental difference between the institutional environment and resource constraints of the two types of enterprises in coping with external shocks. By virtue of their institutional advantages, SOEs are better able to buffer supply chain risk shocks. On the one hand, implicit guarantees and policy support make it easier for them to obtain government bailouts, policy loans, or tax breaks when experiencing supply chain disruptions caused by climate shocks, thus reducing their reliance on internal cash; on the other hand, SOEs have stronger access to resources, which can be deployed through the channels of energy, logistics, and key raw materials to reduce coping costs. In addition, soft budget constraints make their precautionary saving incentives relatively weak and their cash holding adjustments small. In contrast, non-SOEs lack institutional protection and rely more on internal cash in the face of supply chain climate shocks, leading to a significant decline in their cash holding levels. Thus, the institutional advantages of SOEs act as a “cushion” to a certain extent, while the cash holdings of non-SOEs are more vulnerable to supply chain climate risks.

From the perspective of financing constraints, there are significant differences in the impact of supply chain climate risks on enterprises’ cash holdings. Empirical results show that enterprises with high financing constraints are significantly affected by shocks, while those with low financing constraints are not significantly affected. This is because enterprises with high financing constraints have insufficient availability of external financing, making it difficult to quickly replenish liquidity through bank, bond or equity financing. They can only rely on internal cash as a buffer. Secondly, the characteristic of “cash is king” makes cash the only safety cushion for dealing with uncertainties and risks. Once supply chain risks occur, enterprises need to quickly use cash to maintain operations. Furthermore, financing constraints amplify the effect of cash consumption. Enterprises with high financing constraints lack stable credit support. When facing supply disruptions or rising costs, their cash reserves will be rapidly depleted. In contrast, enterprises with low financing constraints have stronger external financing capabilities, can obtain liquidity support through diversified channels, do not need to extensively use internal cash, and can even maintain a relatively high cash reserve to cope with risk shocks. In conclusion, financing constraints have significantly enhanced the negative impact of supply chain climate risks on cash holdings. For enterprises with high financing constraints, due to limited external funds, the decline in cash holdings is more pronounced.

With regard to bank-enterprise relationship, the closeness of cooperation between enterprises and banks significantly affects the role of supply chain climate risk on cash holdings, and the empirical results in

Table 8. Heterogeneity analyses—Based on bank–firm relationships, and the level of regional financial development.

	(1)	(2)	(3)	(4)
	CashHolding (Without bank–firm relationships)	CashHolding With bank–firm relationships	CashHolding (Regional finance developing)	CashHolding (Regional finance developed)
CR	−0.7823 *	−0.2895	−0.8887 **	−0.2719
	(0.4401)	(0.2112)	(0.3446)	(0.3245)
_cons	−0.2735	−0.0398 **	−0.0667 ***	−0.0541
	(0.2573)	(0.0157)	(0.0250)	(0.1065)
Controls	Yes	Yes	Yes	Yes
Firm FE	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes	Yes
N	5509	9036	6630	6594
R2	0.4931	0.5093	0.5784	0.3315

Note: The report includes robust standard errors that are clustered at the firm level, indicated in brackets. *** p < 0.01; ** p < 0.05; * p < 0.1.

show that enterprises with bank-enterprise relationship are not significantly affected by shocks, while enterprises without bank-enterprise relationship are subjected to significant negative shocks. Firstly, this is because the trust and information symmetry established through long-term cooperation between firms and banks allows banks to have a good understanding of the operating conditions of firms, and when supply chain shocks occur, this trust can help firms to obtain emergency loans or expand credit lines quickly without lengthy approvals. Secondly, flexible financing channels provide additional liquidity protection, and the bank-enterprise relationship enables enterprises to have a reliable external “cash pool”, which can be utilized in a timely manner to reduce the reliance on internal cash. Thirdly, the insurance effect strengthens the confidence of enterprises. A good bank-corporate relationship is similar to “insurance”, which enables enterprises to rely on bank support to cushion the impact when risks occur. In sum, as a micro manifestation of external financing constraints, bank-corporate relationship effectively alleviates the pressure of supply chain climate risk on cash holding; while firms without bank-corporate relationship rely more on internal cash, which puts more pressure on coping with the risk.

When it comes to the level of local financial development, the financial environment of the enterprise’s location significantly affects the role of supply chain climate risk on cash holdings, and the empirical results show that enterprises located in financially developed regions are not significantly affected by shocks, while the impact of enterprises in less financially developed regions is more significant. This is due to the fact that diversified financing channels provide more external sources of funds, and financially developed regions have a well-developed banking system, an active capital market, and a wealth of non-bank financial institutions, so that firms can quickly obtain liquidity support when supply chain shocks occur, and reduce their reliance on internal cash. Secondly, competitive credit market reduces financing costs and improves service efficiency, and enterprises can quickly choose the optimal financing solution to cope with unexpected financial needs. Again, a favorable financial ecosystem, including a sound credit system, rule of law, and transparent information disclosure, reduces financing transaction costs and makes it easier for external funds to replace internal cash reserves. To summarize, the level of local financial development determines the advantages and disadvantages of the external financing environment of enterprises at the macro level, and enterprises in developed regions have more stable cash holdings, while enterprises in less developed regions rely more on internal cash and cope with greater risk pressure.

Combining the four heterogeneity analyses reveals that the financial impact of supply chain climate risk on firms’ cash holdings depends on the “financial resilience” of firms to external shocks. The nature of firms’ ownership, financing constraints, banking relationships and the level of local financial development together shape their ability to cope with supply chain climate risk and determine the sensitivity of cash holdings to external shocks.

6. Conclusions

Against the backdrop of intensifying global climate change and elevated external uncertainty, how firms can effectively manage financial risks in supply chains has become a central concern for managers and policy makers. Using a sample of Chinese A-share listed companies, this paper investigates the impact of supply chain climate risk on firms’ cash holdings and further explores the moderating role of supplier concentration and AI application. The empirical results show that supply chain climate risk significantly reduces the level of firms’ cash holdings, supporting the hypothesis that firms face reduced cash holdings in response to supply chain climate shocks. However, firms with high supplier concentration are able to anticipate risks more effectively and increase cash reserves in a timely manner. In addition, the application of AI, especially using Key AI, can mitigate the negative impact of climate risk on cash holdings and enhance firms’ coping ability by improving risk identification and resource allocation efficiency. This study not only verifies the impact of supply chain climate risk on corporate cash holdings, but more importantly proposes an integrated framework of “external climate shock-organizational buffer-AI-enabled”, which organically integrates climate risk, supply chain management and corporate financial behavior. Unlike previous literature that focuses on macro climate risk or internal governance factors, this study emphasizes the heterogeneous mechanisms of firms’ responses to environmental shocks at the supply chain level and reveals the key role of AI technology in enhancing risk management capabilities. This integrative framework not only deepens the understanding of the logic of firms’ financial decision-making under environmental uncertainty but also provides a new theoretical path for future research at the intersection of climate risk and firm behavior.

Also, these findings have important theoretical and practical implications. Theoretically, the study enriches the corporate financial management literature by linking supply chain climate risk, corporate relationship networks, and technology applications to cash holding behavior. In practice, firms should assess supply chain climate risk and optimize their cash management strategies by combining supplier concentration and AI technology capabilities. Enterprises can improve financial resilience by establishing robust supply chain relationships, strategic bank-enterprise relationships, optimizing financing structures, and expanding diversified financing channels. At the same time, they should improve internal cash management to ensure sufficient liquidity despite external financing constraints, and utilize AI-powered supply chain monitoring systems to analyze supplier and climate data in real time to improve risk prediction and liquidity planning capabilities. Policymakers should support enterprises, especially private and small and medium-sized enterprises, to build risk monitoring systems and promote the implementation of AI technology, as well as optimize the external financing environment, including the provision of credit guidance, financing guarantees, tax incentives, venture capital and other policy support, and enhance the level of regional financial development and the efficiency of the credit market, so as to cushion the financial pressure of the enterprises due to the supply chain climate shock. By deepening the market-oriented reform of the financial sector and improving the green finance system, when climate shocks occur, finance can be more efficiently allocated to enterprises with strong adaptive capacity and potential for sustainable development, thereby enhancing the overall climate resilience of the economic system.

Nonetheless, this study has some limitations. First, the examination of supply chain climate risk in this paper focuses only on the supplier level and has not yet explored the potential counterproductive effects of client-side climate risk on firms’ financial performance. Second, the supply chain climate risk indicators constructed in this study only cover domestic supply chain networks and do not include the impact of international supply chain networks. However, these limitations themselves point to promising directions for future research. It would be of great theoretical value to place the supply chain climate risk framework in the context of cross-border supply networks and to systematically assess the interactive effects of carbon policies, trade regimes and climate risk in different countries on the financial behavior of enterprises. Meanwhile, future research can also explore the two-way transmission mechanism of “supplier risk–customer risk” and construct a more comprehensive model of the financial effects of climate risk. Therefore, the findings of this study not only contribute to the current academic discussion but also lay the foundation for future research on climate resilience of multinational and multi-node supply chains.

In addition, this study is important for promoting sustainable development. This paper reveals how firms can enhance financial and operational resilience under supply chain climate shocks, providing a quantitative analytical basis for building more climate resilient and financially sustainable firms. The study further finds that supplier concentration and artificial intelligence applications can improve risk control and dynamic adjustment capabilities, mitigate environmentally induced financial vulnerability, and thus enhance firms’ sustainable operations in the face of external uncertainty. From the policy level, the findings of this study suggest that the organic integration of climate risk management and digital transformation not only enhances sustainability at the firm level, but also helps to promote the transition of the macroeconomic system towards a more resilient and sustainable direction.