Weathering the storm: climate risk perception and digital innovation in firms with exposure to the grain sector

Abstract

Introduction:

Global climate change has increased the frequency of extreme weather events, making climate risk a salient external shock to firms’ sustainable operations. Firms with exposure to the grain sector, as key actors in the food security system, rely heavily on natural conditions and therefore tend to be more sensitive to climate variability.

Methods:

Using panel data on A-share listed firms in China with exposure to the grain sector from 2016 to 2024, this paper employs a fixed-effects model to examine the effect of climate risk perception on digital innovation in firms with exposure to the grain sector and its underlying channels.

Results:

Climate risk perception has a positive and statistically significant effect on digital innovation among these firms, and the estimates are robust to a range of alternative specifications. Mechanism tests suggest that the effect operates primarily through increased R&D investment and improvements in human capital. Moderation analysis further shows that managerial myopia attenuates, whereas organizational responsiveness strengthens, the positive relationship between climate risk perception and digital innovation in firms with exposure to the grain sector.

Discussion:

These findings offer theoretical and empirical insights for emerging market economies aiming to manage climate risk and enhance food security.

1 Introduction

Extreme weather events caused by global climate change are becoming more frequent and intense, climate risk has become a significant external shock that threatens the sustainability of businesses (Tian et al., 2025). In climate economics, these risks are commonly categorized as physical and transition risks (Stroebel and Wurgler, 2021). Physical risks arise from direct impacts such as extreme weather events, sea level rise, and natural disasters, transition risks stem from policy reforms, technological change, and shifts in market preferences as the global economy moves toward a low carbon model (Bao and Zhang, 2026). A growing body of research examines the effects of climate change on macroeconomic performance, financial stability, and firm behavior, highlighting the pervasive and systemic nature of climate risk in modern economic and business environments (Giglio et al., 2021).

The widespread diffusion of digital technologies has stimulated a growing and increasingly sophisticated body of research on digital innovation. Yoo et al. (2010) were among the first to conceptualize digital innovation as the recombination of digital and physical components that generates new products, services, and business models, capturing the innovative outcomes of digital technology adoption. As key actors in the food security system, firms with exposure to the grain sector depend heavily on natural conditions and are therefore particularly vulnerable to climate fluctuations (Yoo et al., 2012). Digital technologies, characterized by reprogrammability and generativity, have become strategic resources in the digital economy (Giustiziero et al., 2023). Such resources can generate sustained competitive advantages and support the long term development of grain related firms (Barney, 1991).

Digital technological innovation can transcend organizational and industry boundaries, enabling firms to adapt to external shocks, sustain competitive advantages, and achieve long term development (Nambisan et al., 2017). Dynamic capability theory posits that firms must develop the capacity to sense opportunities and threats, seize them, and reconfigure resources in order to sustain innovation under uncertainty. In the context of climate change, firms with exposure to the grain sector face increasingly frequent extreme weather events, policy adjustments, and market shifts (Calvet et al., 2022). Climate risk has therefore become a major source of strategic uncertainty. Firms’ perceptions of climate risk constitute the starting point for activating dynamic capabilities. Heightened risk awareness may encourage greater investment in digital technologies and a reallocation of research and development toward risk mitigation and new opportunities. At the same time, risk aversion may crowd out long term innovation investment. Against this backdrop, examining the relationship between climate risk perception and digital innovation in firms with exposure to the grain sector has significant theoretical and practical relevance. Deeper understanding of this relationship can inform firms’ responses to climate risk, advance high quality development in the grain sector, and reinforce national food security. Accordingly, this article seeks to address three questions. Does climate risk perception affect corporate digital innovation in firms with exposure to the grain sector? Through what channels does this effect operate? Does the impact exhibit heterogeneity across firms or contexts? Using a sample of Chinese firms with exposure to the grain sector, this paper examines the relationship between climate risk perception and digital innovation at the firm level. Our contributions can be summarized along three dimensions.

Firstly, this paper offers a new perspective by jointly examining climate risk perception and firm-level digital innovation in Chinese firms with exposure to the grain sector. Rather than focusing on a single dimension, the analysis integrates these two strands of inquiry into a unified empirical framework. This approach contributes to the emerging literature at the intersection of climate adaptation and the digital economy, particularly in this understudied setting.

Secondly, this article identifies two main channels through which climate risk perception influences digital innovation among Chinese firms with exposure to the grain sector: R&D investment and human capital upgrading. By clarifying these channels, our analysis sheds light on the underlying transmission process and extends a literature that has primarily documented empirical regularities, often without rigorous causal identification.

Thirdly, in terms of methodology, this paper not only provides a detailed framework for identifying firms with exposure to the grain sector, but also constructs a climate risk perception measure using text-based word frequency analysis. Based on this, a fixed effects model is employed to assess the impact of climate risk perception on digital innovation in these firms. This approach offers valuable methodological insights for future research and provides empirical support for both practice and policy formulation in China’s grain sector.

2 Literature review and theoretical framework

2.1 Literature review

2.1.1 The research of climate risk perception

In definitional terms, climate risk perception refers to firms’ awareness and understanding of potential extreme climate events and the associated threats and losses these events may pose to their operations (Tian et al., 2024). In terms of measurement, existing studies typically proxy climate risk using losses caused by extreme weather events, the intensity of El Niño and La Niña episodes, and indicators of climate policy uncertainty (Ling and Gao, 2023; Liu et al., 2024; Qin et al., 2024). In recent years, some studies have used text analysis to measure firms’ climate risk perception. By examining climate-related words in annual reports and other public disclosures, these studies offer an intuitive picture of firms’ awareness of climate risk (Li et al., 2024). A growing body of research uses text analysis to examine how climate risk shapes firms’ ESG performance, investment efficiency, risk management, and strategic choices (Berkman et al., 2024; Duan et al., 2025; Zhao and Li, 2026). These studies highlight the critical importance of actively addressing climate risk to ensure long-term sustainability. Firms, as the core drivers of economic development, play a crucial role in addressing and mitigating the risks of climate change while advancing sustainable socio-economic growth (Andersson et al., 2016).

2.1.2 Determinants of digital innovation

A growing body of research examines digital innovation at the firm level, typically defining it as the integration of digital technologies, such as artificial intelligence, blockchain, cloud computing, and big data, with physical operations (Goebeler et al., 2024). This process encompasses product design optimization, production transformation, and business model innovation (Ganotakis et al., 2023). Digital innovation has become a central strategy for firms seeking competitive advantage. The literature has made substantial progress in identifying the factors that drive or constrain firm-level digital innovation. Regarding internal factors, sufficient R&D funding ensures the smooth progress of digital innovation (Chu et al., 2021), while highly skilled, multidisciplinary talent provides strategic guidance for innovation (Deist et al., 2023). On the external front, government data governance and digital economy policies play a crucial role in incentivizing firms to pursue digital innovation (Xiao et al., 2024).

2.1.3 Climate risk and digital innovation

Climate risk increases operational uncertainty and disrupts production and supply chains. Such pressures heighten demand for timely information processing, risk forecasting, and resource reallocation. Digital technologies thus serve as a key strategic tool by strengthening monitoring, coordination, and decision-making capabilities (Chen and Zhang, 2025; Zhao and Liao, 2025). Existing research shows that climate risk induces firms to adopt adaptive strategies, pursue green transformation, and upgrade technologies, and digital technologies help transform environmental uncertainty into manageable operational risks (Jin and Gao, 2025; Yin et al., 2024). This body of work highlights the close interconnection between climate risk and digital innovation.

Overall, prior research has focused either on the determinants of digital innovation or on the economic effects of climate risk (Ciarli et al., 2021; Giglio et al., 2021; Liu et al., 2023). The intersection of climate risk perception and digital innovation in firms with exposure to the grain sector remains largely unexplored. In particular, evidence on the underlying linkages and transmission mechanisms is still limited.

2.2 Theoretical framework

2.2.1 Climate risk perception and digital innovation in firms with exposure to the grain sector

Climate risk constitutes an objective external shock arising from climate change. Firms’ climate risk perception captures managers’ recognition, interpretation, and assessment of this risk. In practice, strategic decisions depend not only on the risk itself but also on firms’ perceptions and responses to it. From a dynamic capability perspective, perception initiates the response process. Recognition of the strategic importance of climate risk prompts firms to pursue new forms of adaptation, reallocate resources, and build capabilities. From a risk management perspective, risk perception encourages preventive and adaptive actions, as managers seek to reduce uncertainty through more informed and disciplined decisions. For firms exposed to the grain sector, this mechanism is particularly pronounced, as production, storage, logistics, and supply stability depend heavily on weather conditions and seasonal factors. This high level of exposure heightens firms’ awareness of climate risks and drives them to transform these challenges into opportunities for innovation. By fully integrating digital technologies, firms can achieve precision agriculture, smart storage, real-time logistics tracking, and comprehensive quality traceability. Accordingly, this article proposes the following testable hypothesis.

H1: Ceteris paribus, climate risk perception increases digital innovation in firms with exposure to the grain sector.

2.2.2 The R&D investment channel

Climate risk perception directly drives digital innovation and may also promote it indirectly by stimulating increased R&D investment. Digital innovation in firms typically requires substantial investment and extended development cycles, depending on continuous funding, advanced equipment, and targeted R&D projects. Climate risk perception, as a central cognitive process that helps firms recognize external threats and anticipate operational uncertainties, directly influences the scale, allocation, and focus of R&D investment. When firms perceive rising climate risks, they tend to increase investment in both fixed assets and R&D (Tushman and Romanelli, 1983). Acknowledging that traditional technologies cannot adequately address climate-driven operational risks, firms with exposure to the grain sector actively expand R&D efforts in drought-resistant crop varieties, smart irrigation systems, pest and disease early-warning platforms, and IoT-enabled grain storage. These activities form a core component of digital innovation. Simultaneously, accumulated R&D builds the knowledge base and talent pool needed for subsequent digital technology applications, establishing a reinforcing cycle of risk perception, R&D investment, and digital innovation. Building on this, the study proposes the following testable hypothesis.

H2: Ceteris paribus, climate risk perception promotes digital innovation in firms with exposure to the grain sector by increasing R&D investment.

2.2.3 The human capital channel

Digital innovation in firms relies on human capital with expertise and skills in digital technologies (Bendig et al., 2023). Human capital theory suggests that a firm’s innovative capacity relies fundamentally on the creative contributions of highly skilled personnel. Digital innovation, in particular, depends on the support of versatile digital talent. When firms perceive rising climate risk, they increase R&D investment and enhance the recruitment and development of digital talent (Dutta and Naveen, 2026). The influence of climate risk perception on human capital can be examined from both the demand and supply perspectives. From the demand side, climate risk perception increases firms’ recognition of the urgency for digital transformation, leading to a heightened demand for digital talent, including data analysts, algorithm engineers, and agricultural IT specialists (Cao et al., 2025). On the supply side, growing climate risk alters the labor market’s skill demands. Digital talent commands a higher market premium due to its ability to help firms navigate environmental uncertainties.

As a result, firms are incentivized to allocate more resources to talent acquisition and skill development. In firms with exposure to the grain sector, human capital plays a particularly crucial role. The food supply chain spans multiple stages, including pre-production research, planting and harvesting, storage and transportation, and processing and sales. Each stage of digital transformation relies on specialized talent. Precision agriculture requires professionals skilled in both agronomy and remote sensing, while smart grain storage facilities demand experts proficient in IoT devices and grain preservation techniques. Firms with exposure to the grain sector that perceive high climate risk actively recruit and develop versatile talent, leveraging accumulated human capital to support digital innovation. Based on this, we propose the following testable hypothesis.

H3: Ceteris paribus, climate risk perception affects digital innovation in firms with exposure to the grain sector through human capital upgrading.

3 Empirical strategy, variables and data

3.1 Empirical strategy

To examine the effect of climate risk perception on digital innovation in firms with exposure to the grain sector, this article estimates the fixed-effects model specified in Equation 1.where i represents the firm, p represents the industry, and t denotes the year. lnDP_ipt represents digital innovation in firms with exposure to the grain sector i in year t. lnCRA captures climate risk perception for firm i in year t. β measures the impact of climate risk perception on digital innovation in firms with exposure to the grain sector. Controls represents a set of control variables. u_p denotes industry fixed effects, controlling for time-invariant factors at the industry level that may affect the regression results. λ_t represents year fixed effects, controlling for time-related factors. ε_ipt is the random error term.

3.2 Variables

3.2.1 Dependent variable: corporate digital innovation

Building on Tan and Cui (2026), this study examines emerging digital industries and frontier technologies. Following the Classification System for Key Digital Technology Patents (2023),¹ IPC codes identify seven categories of key digital technologies: artificial intelligence, advanced chips, quantum information, the IoT, blockchain, the industrial internet, and the metaverse. Firms’ digital technologies are typically application oriented and evolve more rapidly than conventional technologies. In this setting, patent applications more effectively capture innovation activity in digital technologies, reflecting firms’ responsiveness to new opportunities. Exclusive reliance on patent applications, however, may capture low-quality patent inflation. To provide a more comprehensive measure of digital innovation, this study follows Qiu et al. (2026) and employs both the logarithm of key digital technology patent applications (lnDPA) and the logarithm of granted patents (lnDPG). Log transformations of both variables mitigate right skewness in patent counts and reduce its influence on the regression results.

3.2.2 Core independent variable: climate risk perception

From a theoretical perspective, recent studies draw on firms’ annual reports to capture perceptions of external risk. These studies examine both regulatory and market risk (Feng et al., 2023; Kölbel et al., 2024). This paper constructs a measure of climate risk perception in several steps. First, annual reports of all A-share listed firms with exposure to the grain sector over the sample period are collected. Second, all firm reports and materials from the China National Meteorological Science Data Center are reviewed, with attention to the Chinese context and language. Follows Li et al. (2024) text analysis and machine learning are used to train the corpus. A climate risk dictionary with extensive feature terms is developed using the Continuous Bag-of-Words model, as reported in Table A1. This paper constructs a climate risk perception index for A-share listed firms, using the natural logarithm of the frequency of climate-related terms as the key explanatory variable (lnCRA). This article adopts text analysis for two main reasons. On one hand, the approach captures firms’ subjective perceptions of climate risk rather than relying solely on objective shocks or external climate indicators. On the other hand, annual reports offer firm-year measures of climate risk exposure that better align with the managerial decision-making mechanisms examined in this study.

3.2.3 Control variables

To reduce omitted variable bias in the regression, this paper follows Tan and Cui (2026) in selecting a set of control variables. Firm size (lnSAL) is the natural logarithm of total revenue. Firm age (lnAGE) is the natural logarithm of the difference between the current year and the founding year. Leverage (ALR) is the ratio of total liabilities to total assets. Cash flow (CAS) is the ratio of operating cash flow to total assets. Fixed asset ratio (FIX) is measured as the ratio of fixed assets to total assets. Return on assets (ROA) is the ratio of total profit to total assets. Ownership concentration (OC) is the ratio of shares held by the top 10 shareholders to total equity. Definitions and descriptions of all variables are presented in Table 1.

Table 2 presents the descriptive statistics of the main variables. lnDPA and lnDPG have means of 0.682 and 0.535, with standard deviations of 1.037 and 0.901, indicating substantial variation in digital innovation among Chinese firms with exposure to the grain sector. lnCRA has a mean of 4.606, a maximum of 6.397, a minimum of 2.565, and a standard deviation of 0.684, highlighting significant differences in climate risk perception among the sample firms during the study period. The considerable variation in these key variables provides a strong basis for econometric identification of causal effects.

3.3 Data sources and processing methods

The sample consists of A-share listed firms in China with exposure to the grain sector over the period 2016 to 2024. Firm-level data are obtained from the China Research Data Services Platform (CNRDS) and the China Stock Market and Accounting Research (CSMAR) database. Identifying firms with exposure to the grain sector is central to the analysis. The procedure proceeds in two steps. Firstly, we select listed firms classified in agriculture, forestry, animal husbandry, fisheries, agricultural and sideline food processing, food manufacturing, wholesale, transportation, and warehousing. Secondly, we manually review firms’ business scope and primary activities to determine whether they have substantive exposure to the grain sector. Firms in the alcohol, beverage, and refined tea manufacturing industries are excluded. Although these firms use grain as an input, their core activities are not directly related to food security, which is the focus of this study. Two examples clarify the identification of firms with exposure to the grain sector. Firms whose core business includes edible vegetable oil production are classified as firms with exposure to the grain sector, as this activity lies at the core of the grain industrial chain. Retail firms are generally excluded, since grain-related activities rarely constitute their main business and grain policies seldom affect the retail segment. The data are further cleaned as follows. Firms under abnormal operating status are excluded, and observations with missing values for key variables are dropped. To mitigate the influence of outliers, all continuous variables are winsorized at the 1% level in both tails.

4 Empirical results

4.1 Baseline regression results

Table 3 reports the baseline regression results on the effect of climate risk perception on digital innovation in firms with exposure to the grain sector. Columns (1) and (2) exclude control variables, controlling only for industry and year fixed effects, with robust standard errors clustered at the industry-year level. Columns (3) and (4) include all control variables. As shown in Table 3, columns (1) to (4), the regression coefficient for lnCRA remains significantly positive at the 1% level. This suggests that climate risk perception effectively drives digital innovation in firms with exposure to the grain sector, as evidenced by increases in both digital patent applications (lnDPA) and granted patents (lnDPG). The estimates in columns (3) and (4) imply economically meaningful effects. Holding other variables constant, a 1% increase in climate risk perception is associated with a 0.221% increase in digital patent applications and a 0.197% increase in granted digital patents, indicating a substantial economic effect.

4.2 Robustness tests

4.2.1 Endogeneity concerns and identification

This paper addresses potential endogeneity in several ways. Firstly, all regressions include industry and year fixed effects. Industry fixed effects account for time-invariant omitted variables at the industry level, while year fixed effects control for time-specific factors. Including control variables further reduces concerns about omitted variables. Secondly, to address potential reverse causality, the specification replaces the explanatory variable with its one-period lag (L. lnCRA). Using lagged core explanatory variables helps rule out reverse causality, since current firm digital innovation cannot affect prior-period climate risk perception. Columns (1) and (2) in Table 4 show that the coefficient on L. lnCRA remains significantly positive.

Thirdly, high-dimensional fixed effects are incorporated to further reduce the potential influence of omitted variables on the regression results. Columns (3) and (4) of Table 4 include province fixed effects to control for time-invariant factors at the province level. Columns (5) and (6) further incorporate province × year fixed effects. This specification absorbs all province-level variation, accounting for both time-invariant and time-varying factors, and represents a highly stringent fixed-effects design. Columns (1) and (2) of Table 5 extend the baseline specification by adding industry-by-year fixed effects, which absorb industry-level omitted factors. Columns (3) and (4) further include province by year and industry by year fixed effects. This specification controls for observable and unobservable factors at the province and industry levels and reduces bias from omitted variables. Across all specifications, the coefficient on lnCRA remains positive and statistically significant at the 1% level, reinforcing the main conclusion.

4.2.2 Additional robustness tests

Firstly, the core explanatory variable is replaced with CRB, defined as the ratio of climate risk related keywords to the total word count in the annual report, and the model is re estimated. Columns (1) and (2) of Table 6 indicate that the coefficient on CRB remains positive and statistically significant at the 1% level, further supporting the main conclusion.

Secondly, robust standard errors are clustered at alternative levels. Because statistical significance may vary with the clustering level, robustness is assessed by clustering the OLS standard errors at the firm and provincial levels, which impose more stringent criteria. Columns (3) to (6) of Table 6 show that the coefficient on lnCRA remains positive and statistically significant under both clustering schemes, consistent with the baseline results.

Overall, the baseline results remain robust after addressing potential endogeneity and conducting additional robustness tests. The evidence indicates that climate risk perception exerts a positive firm-level effect and promotes digital innovation in firms with exposure to the grain sector, consistent with H1.

4.3 Mechanism tests

Following Bai et al. (2025) and Huang et al. (2025), this paper specifies the empirical model in Equation 2 to examine the underlying mechanisms.where M_ipt denotes a set of mechanism variables capturing R&D investment and human capital structure. Details of variable construction are provided below. All other variables follow the specification in Equation 1.

4.3.1 R&D investment channels

To examine whether climate risk perception affects corporate digital innovation through the R&D channel, this paper considers two dimensions: R&D expenditure and R&D personnel, using the same controls and baseline specification. R&D funding is measured by the natural log of R&D expenditure (lnRD) and R&D intensity (RDS). R&D personnel is measured by the natural log of R&D staff (lnRP) and their share of total employees (RDP). Each of these four measures is then used as the dependent variable in separate re-estimations of the models. Columns (1) to (4) of Table 7 show that lnCRA coefficients are positive and significant at the 1% level. Climate risk perception therefore increases both the total R&D funding and R&D personnel in firms with exposure to the grain sector, while also raising their respective shares. These findings provide empirical evidence for the R&D investment channel, consistent with H2.

4.3.2 The human capital channel

To test whether climate risk perception influences digital innovation via the human capital channel, we distinguish two education groups: employees with bachelor’s or associate degrees and employees with graduate degrees. For the first group we use the natural log of the number of bachelor’s/associate-degree holders (lnBS) and their share of total employees (BSS). For the second group we use the natural log of the number of employees with graduate degrees (lnMS) and their share of total employees (MSS). We then replace the dependent variable with each of these four measures and re-estimate the empirical models. Columns (5) to (8) of Table 7 indicate that lnCRA is positively and significantly associated with lnBS and lnMS, whereas the coefficients on BSS and MSS are statistically insignificant. Climate risk perception increase the total number of highly educated employees in firms with exposure to the grain sector, but do not significantly alter their share in the short run. These results support the human capital channel, consistent with H3.

4.4 Moderating effects

Having established that climate risk perception enhances digital innovation in firms with exposure to the grain sector, this paper next examines the conditions under which this effect is strengthened. Two moderating factors are considered: managerial myopia and organizational responsiveness. Following Baruffaldi et al. (2024), this study specifies the moderating model in Equation 3 to examine moderating effects.where D_ipt denotes the moderating variables capturing managerial short termism and responsiveness. Details of variable construction are provided below. All other variables follow the specification in Equation 1.

4.4.1 Managerial myopia

Digital innovation is a long-term and inherently uncertain process, and managerial myopia may dampen innovation incentives. Managerial myopia (SH) is measured as the ratio of current short-term investments to beginning-of-period total assets. Columns (1) and (2) of Table 8 report a negative and statistically significant coefficient on lnCRA × SH at the 5% level or better, indicating that managerial myopia attenuates the positive effect of climate risk perception on digital innovation among firms with exposure to the grain sector.

4.4.2 Organizational responsiveness

Organizational responsiveness reflects a firm’s ability to adapt to changing market conditions and drive growth through rapid and innovative actions (Lu and Ramamurthy, 2011). This capability is closely tied to the board of directors, as major decisions in publicly listed companies are generally deliberated and approved at the board level. Higher organizational responsiveness enables firms to respond to perceived climate risks with timely and appropriate decisions. Consequently, it may strengthen the positive relationship between climate risk perception and digital innovation in firms with exposure to the grain sector. This paper tests this logic by measuring organizational responsiveness through the number of board meetings (BMF) and shareholder meetings (GMF) and applying a moderation model for empirical analysis. Columns (3) to (6) of Table 8 show positive and significant coefficients for lnCRA × BMF and lnCRA × GMF. The results indicate that higher organizational responsiveness amplifies the positive effect of climate risk perception on digital innovation in firms with exposure to the grain sector, confirming our hypothesis.

5 Discussion

5.1 Climate risk perception significantly promotes digital innovation in firms with exposure to the grain sector

The results show that climate risk perception significantly promotes digital innovation among firms with exposure to the grain sector. Climate risk thus acts not only as a constraint but also as a catalyst for strategic adjustment. For firms with exposure to the grain sector, short-term cost adjustments fail to mitigate climate shocks effectively. Digital innovation offers a powerful tool to boost adaptability and manage uncertainty (Browder et al., 2024). These findings indicate that severe environmental challenges and food security pressures heighten the responsibility of firms with exposure to the grain sector to safeguard food supplies. This responsibility, in turn, converts climate risk perception into a powerful driver of innovation.

5.2 R&D investment and human capital upgrading as key channels linking climate risk perception to digital innovation in firms with exposure to the grain sector

Increased R&D investment and human capital upgrading form a dual channel through which climate risk perception drive digital innovation in firms with exposure to the grain sector, establishing a complete pathway for transforming resources into innovation. From the R&D perspective, digital innovation demands substantial investment and strong contextual adaptation, making continuous and targeted allocation of R&D resources essential throughout the entire innovation cycle (Rammer et al., 2022). Climate risk perception lead grain-sector-exposed firms to direct R&D resources toward climate-adaptive digital technologies. These activities form a core component of digital innovation and establish the technical foundation for integrating technology with industry-specific applications, generating a positive cycle of risk perception, targeted R&D allocation, and innovation implementation (Lee et al., 2025).

From the human capital perspective, firms respond to climate risk mainly by adjusting labor input rather than rapidly restructuring the workforce. Among firms with exposure to the grain sector, employees with higher education still represent a small share of the workforce. Most positions are held by workers performing routine, traditional tasks. This pattern reflects the sector’s large conventional labor base and rigid talent structure rather than limitations in recruitment channels. When firms perceive increasing climate risks, they often recruit, retain, or upskill highly qualified employees to manage digital-related tasks (Sautner et al., 2023). R&D and human capital work together, creating a full pathway through which climate risk perception promotes digital innovation.

5.3 Climate risk perception has a stronger positive effect on digital innovation in firms with exposure to the grain sector that demonstrate higher organizational responsiveness

Our findings further show that responsiveness in firms with exposure to the grain sector amplifies the positive relationship between climate risk perception and digital innovation. This indicates that risk awareness alone is insufficient; firms must convert perception into prompt action. For firms highly exposed to climate shocks, the sudden and seasonal nature of climate change makes rapid decision-making and flexible resource reallocation crucial for turning risk perception into innovation (Yoo et al., 2012). This implies that the innovative value of climate risk awareness depends on a firm’s ability to respond quickly and reconfigure resources effectively.

6 Conclusions and policy implications

6.1 Conclusion

Using panel data on A-share listed firms in China with exposure to the grain sector from 2016 to 2024, this paper employs a fixed-effects model to examine the effect of climate risk perception on digital innovation in firms with exposure to the grain sector and its underlying channels. Climate risk perception has a positive and statistically significant effect on digital innovation among these firms, and the estimates are robust to a range of alternative specifications. Mechanism tests suggest that the effect operates primarily through increased R&D investment and improvements in human capital. Moderation analysis further shows that managerial myopia attenuates, whereas organizational responsiveness strengthens, the positive relationship between climate risk perception and digital innovation in firms with exposure to the grain sector.

6.2 Policy implications

6.2.1 Strengthening climate risk perception to support digital innovation in firms with exposure to the grain sector

The government should provide specific guidance to fully integrate climate risk management into the strategic planning for the high-quality development of the food industry. This would encourage firms with exposure to the grain sector to include climate risk perception and digital innovation in their long-term strategies. The government should establish an industry-wide climate risk information-sharing platform that integrates data from meteorology, agriculture, and related sectors. This platform would combine multidimensional data, including weather monitoring, grain production, and supply chain fluctuations, to lower the cost of climate risk identification and strengthen managers’ systemic understanding of these risks (Zhou and Richardson-Barlow, 2026). Additionally, outcomes of climate-adaptive digital innovation should be incorporated into the evaluation systems of leading firms with exposure to the grain sector and key food security enterprises. Such measures would establish industry benchmarks, foster a sector-wide consensus on using digital innovation to mitigate climate risks, and fully unlock firms’ innovation potential.

6.2.2 Strengthening factor support systems to increase investment in R&D and human capital

On one hand, the government should increase R&D support for climate-adaptive digital innovation in firms with exposure to the grain sector. Prioritize funding for smart irrigation, IoT-enabled grain storage and supply-chain risk early-warning systems. Encourage private capital participation to ease firms’ R&D financing constraints. On the other hand, strengthen the pipeline of hybrid talent with expertise in both digital technologies and agriculture by supporting platforms that integrate industry, universities, research institutes and firms. Introduce targeted incentives to lower the costs of recruiting and retaining highly educated, high-skilled digital personnel, ensuring a steady supply of human resources for climate-adaptive digital innovation.

6.2.3 Enhancing corporate governance to strengthen agile responses to climate risk in firms with exposure to the grain sector

Firstly, firms with exposure to the grain sector should continue optimizing their internal governance and decision-making processes by removing institutional barriers between climate risk management and digital innovation decisions. This will enhance the overall agility of the organization in responding to challenges. Secondly, given the seasonal nature of food production and the sudden impact of climate shocks, these firms should improve their dynamic climate risk response and decision-making mechanisms. This will reduce decision-making cycles and increase efficiency. Thirdly, the governance evaluation system for publicly listed grain-related companies should be strengthened by integrating metrics such as climate risk response efficiency and the conversion rate of digital innovation outcomes. These changes would encourage firms to develop agile decision-making and resource allocation systems, thereby maximizing the innovation-driving effect of climate risk perception and enhancing supply chain resilience.

7 Contributions, limitations and future prospects

7.1 Contributions

This paper makes three main contributions. Firstly, it extends the literature by examining how climate risk perception influences digital innovation in A-share listed firms engaged in firms with exposure to the grain sector, where climate change is closely linked to food security and firms’ strategic decisions. Secondly, the analysis identifies two key transmission channels, R&D investment and human capital upgrading, and shows that managerial myopia weakens the effect, whereas organizational responsiveness strengthens it. Thirdly, the study develops a measure of climate risk perception and establishes a transparent procedure for identifying risks in the grain sector, providing a replicable empirical framework for future research.

7.2 Limitations and future prospects

Although the findings offer some marginal contributions, this paper still has limitations. On one hand, the sample includes only Chinese A-share listed firms, so the results may not generalize to unlisted firms. Future research could examine similar questions using data from non-listed firms and small and micro enterprises to assess whether the findings generalize to broader settings. On the other hand, although this study addresses endogeneity using lagged variables and high-dimensional fixed effects, these approaches mainly mitigate reverse causality and omitted variable bias separately, rather than jointly. Future research could explore the use of a valid instrument for climate risk perception and apply a 2SLS framework to obtain more credible estimates.

References

• 1

  AnderssonM.BoltonP.SamamaF. (2016). Hedging climate risk. Finance Anal. J.72, 13–32. doi: 10.2469/faj.v72.n3.4

  • CrossRef
  • Google Scholar
  • View reference in article

• 2

  BaiC.JiaR.LiH.WangX. (2025). Entrepreneurial reluctance: talent and firm creation in China. Econ. J.135, 964–981. doi: 10.1093/ej/ueae094

  • CrossRef
  • Google Scholar
  • View reference in article

• 3

  BaoX.ZhangX. (2026). Climate risk perception and corporate green innovation: from cognitive awareness to behavioral response. Sustainability18:752. doi: 10.3390/su18020752

  • CrossRef
  • Google Scholar
  • View reference in article

• 4

  BarneyJ. (1991). Firm resources and sustained competitive advantage. J. Manage.17, 99–120. doi: 10.1177/014920639101700108

  • CrossRef
  • Google Scholar
  • View reference in article

• 5

  BaruffaldiS.SimethM.WehrheimD. (2024). Asymmetric information and R&D disclosure: evidence from scientific publications. Manag. Sci.70, 1052–1069. doi: 10.1287/mnsc.2023.4721

  • CrossRef
  • Google Scholar
  • View reference in article

• 6

  BendigD.WagnerR.PieningE.FoegeJ. N. (2023). Attention to digital innovation: exploring the impact of a chief information officer in the top management team. MIS Q.47, 1487–1516. doi: 10.25300/MISQ/2023/17152

  • CrossRef
  • Google Scholar
  • View reference in article

• 7

  BerkmanH.JonaJ.SoderstromN. (2024). Firm-specific climate risk and market valuation. Account. Organ. Soc.112:101547. doi: 10.1016/j.aos.2024.101547

  • CrossRef
  • Google Scholar
  • View reference in article

• 8

  BrowderR. E.DwyerS. M.KochH. (2024). Upgrading adaptation: how digital transformation promotes organizational resilience. Strateg. Entrep. J.18, 128–164. doi: 10.1002/sej.1483

  • CrossRef
  • Google Scholar
  • View reference in article

• 9

  CalvetL.GianfrateG.UppalR. (2022). The finance of climate change. J. Corp. Finance73:102162. doi: 10.1016/j.jcorpfin.2022.102162

  • CrossRef
  • Google Scholar
  • View reference in article

• 10

  CaoZ.ChenS. X.LeeE.XieS. (2025). Weathering the storm: unravelling the influence of climate change risk exposure on firms’ human capital investment. Br. J. Manage.36, 633–649. doi: 10.1111/1467-8551.12868

  • CrossRef
  • Google Scholar
  • View reference in article

• 11

  ChenW.ZhangQ. (2025). Can corporate climate risk drive digital transformation? Evidence from Chinese heavy-polluting enterprises. Technol. Forecast. Soc. Change212:123990. doi: 10.1016/j.techfore.2025.123990

  • CrossRef
  • Google Scholar
  • View reference in article

• 12

  ChuC.LiY.LiS.JiY. (2021). Uncertainty, venture capital and entrepreneurial enterprise innovation—evidence from companies listed on China's GEM. Pac.-Basin Finance J.68:101576. doi: 10.1016/j.pacfin.2021.101576

  • CrossRef
  • Google Scholar
  • View reference in article

• 13

  CiarliT.KenneyM.MassiniS.PiscitelloL. (2021). Digital technologies, innovation, and skills: emerging trajectories and challenges. Res. Policy50:104289. doi: 10.1016/j.respol.2021.104289

  • CrossRef
  • Google Scholar
  • View reference in article

• 14

  DeistM. K.McdowellW. C.BounckenR. B. (2023). Digital units and digital innovation: balancing fluidity and stability for the creation, conversion, and dissemination of sticky knowledge. J. Bus. Res.161:113827. doi: 10.1016/j.jbusres.2023.113827

  • CrossRef
  • Google Scholar
  • View reference in article

• 15

  DuanD.WeiR.WangC.XiaB. (2025). Opportunity or obstacle? Climate risk disclosure and corporate ESG performance. Int. Rev. Econ. Finance100:104101. doi: 10.1016/j.iref.2025.104101

  • CrossRef
  • Google Scholar
  • View reference in article

• 16

  FengY.LiJ.SunX.LiG. (2023). Do companies' risk perceptions affect investor confidence? Evidence from textual risk disclosure in the tourism industry. Tour. Manag. Perspect.49:101189. doi: 10.1016/j.tmp.2023.101189

  • CrossRef
  • Google Scholar

• 17

  GanotakisP.AngelidouS.SaridakisC.PiperopoulosP.DindiaM. (2023). Innovation, digital technologies, and sales growth during exogenous shocks. Technol. Forecast. Soc. Change193:122656. doi: 10.1016/j.techfore.2023.122656

  • CrossRef
  • Google Scholar

• 18

  GiglioS.MaggioriM.RaoK.StroebelJ.WeberA. (2021). Climate change and long-run discount rates: evidence from real estate. Rev. Finance Stud.34, 3527–3571. doi: 10.1093/rfs/hhab032

  • CrossRef
  • Google Scholar

• 19

  GiustizieroG.KretschmerT.SomayaD.WuB. (2023). Hyperspecialization and hyperscaling: a resource-based theory of the digital firm. Strategic Manage. J.44, 1391–1424. doi: 10.1002/smj.3365

  • CrossRef
  • Google Scholar

• 20

  GoebelerL.HukalP.XiaoX. (2024). Four roles of physicality in digital innovation: a theoretical review. J. Strateg. Inf. Syst.33, 1–22. doi: 10.1016/j.jsis.2024.101862

  • CrossRef
  • Google Scholar

• 21

  DuttaD.NaveenP. M. (2026). Transforming recruitment and selection practices in organizations through discriminative and generative AI adoption: a structuration lens. Hum. Resour. Manag.65, 77–115. doi: 10.1002/hrm.70018

  • CrossRef
  • Google Scholar

• 22

  HuangW.LiuQ.XingJ.ZhengS. (2025). Breaking the iron rice bowl: tenure-track systems and the rise of academic innovation in China. Econ. J.:ueaf093. doi: 10.1093/ej/ueaf093

  • CrossRef
  • Google Scholar

• 23

  JinJ.GaoB. (2025). Going with the flow or standing by: managerial climate risk perception bias and corporate green transformation — evidence from China. Energy Econ.148:108694. doi: 10.1016/j.eneco.2025.108694

  • CrossRef
  • Google Scholar

• 24

  KölbelJ. F.LeippoldM.RillaertsJ.WangQ. (2024). Ask BERT: how regulatory disclosure of transition and physical climate risks affects the CDS term structure. J. Finance Econ.22, 30–69. doi: 10.1093/jjfinec/nbac027

  • CrossRef
  • Google Scholar

• 25

  LeeC.LiM.ZhangJ. (2025). How climate risk affects corporate green innovation: fresh evidence from China's listed companies. Emerg. Mark. Finance Trade61, 2302–2315. doi: 10.1080/1540496X.2024.2449463

  • CrossRef
  • Google Scholar

• 26

  LiQ.ShanH.TangY.YaoV. (2024). Corporate climate risk: measurements and responses. Rev. Finance Stud.37, 1178–1830. doi: 10.1093/rfs/hhad094

  • CrossRef
  • Google Scholar

• 27

  LingS.GaoH. (2023). How does climate risk matter for corporate green innovation? Empirical evidence from heavy-polluting listed companies in China. Front. Energy Res.11:1177927.

  • Google Scholar

• 28

  LiuY.DongJ.MeiL.ShenR. (2023). Digital innovation and performance of manufacturing firms: an affordance perspective. Technovation119:102458. doi: 10.1016/j.technovation.2022.102458

  • CrossRef
  • Google Scholar

• 29

  LiuX.WuX.KongX. (2024). Do climate-related risks perception drive corporate green and low-carbon transformation? Evidence from listed companies in China. Emerging Mark. Finance Trade60, 2938–2959. doi: 10.1080/1540496X.2024.2345190

  • CrossRef
  • Google Scholar

• 30

  LuY.RamamurthyK. (2011). Understanding the link between information technology capability and organizational agility: an empirical examination. MIS Q.35, 931–954. doi: 10.2307/41409967

  • CrossRef
  • Google Scholar

• 31

  NambisanS.LyytinenK.MajchrzakA.SongM. (2017). Digital innovation management: reinventing innovation management research in a digital world. MIS Q.41, 223–238. doi: 10.25300/MISQ/2017/41:1.03

  • CrossRef
  • Google Scholar

• 32

  QinM.ZhuY.XieX.ShaoX.LonbontO. (2024). The impact of climate risk on technological progress under the fourth industrial era. Technol. Forecast. Soc. Change202:123325. doi: 10.1016/j.techfore.2024.123325

  • CrossRef
  • Google Scholar

• 33

  QiuL.WeiZ.HeJ. (2026). How do national wetland parks promote corporate green innovation? Evidence from a quasi-natural experiment in China. J. Clean. Prod.541:147526. doi: 10.1016/j.jclepro.2026.147526

  • CrossRef
  • Google Scholar

• 34

  RammerC.FernándezG. P.CzarnitzkiD. (2022). Artificial intelligence and industrial innovation: evidence from German firm-level data. Res. Policy51:104555. doi: 10.1016/j.respol.2022.104555

  • CrossRef
  • Google Scholar

• 35

  SautnerZ.Van LentL.VilkovG.ZhangR. (2023). Firm-level climate change exposure. J. Finance78, 1449–1498. doi: 10.1111/jofi.13219

  • CrossRef
  • Google Scholar

• 36

  StroebelJ.WurglerJ. (2021). What do you think about climate finance?J. Finance Econ.142, 487–498. doi: 10.1016/j.jfineco.2021.08.004

  • CrossRef
  • Google Scholar

• 37

  TanQ.CuiY. (2026). Can digital innovation be a new driver of corporate sustainability?China J. Account. Res.19:100466. doi: 10.1016/j.cjar.2025.100466

  • CrossRef
  • Google Scholar

• 38

  TianY.ChenS.DaiL. (2024). How climate risk drives corporate green innovation: evidence from China. Finance Res. Lett.59:104762. doi: 10.1016/j.frl.2023.104762

  • CrossRef
  • Google Scholar

• 39

  TianH.WangD.GuanY.LiX. (2025). Climate risk, social resilience and crime rates. Energy Econ.152:109023. doi: 10.1016/j.eneco.2025.109023

  • CrossRef
  • Google Scholar

• 40

  TushmanM. L.RomanelliE. (1983). Uncertainty, social location and influence in decision making: a sociometric analysis. Manag. Sci.29, 12–23. doi: 10.1287/mnsc.29.1.12

  • CrossRef
  • Google Scholar

• 41

  XiaoD.XuJ.LiQ. (2024). The “double-edged sword” effect of air quality information disclosure policy—empirical evidence based on the digital transformation of Chinese listed companies. Energy Econ.133:107513. doi: 10.1016/j.eneco.2024.107513

  • CrossRef
  • Google Scholar

• 42

  YinQ.WangY.SongD.LaiF.CollinsB.GuoH. (2024). The impact of digitalization on operational risk: an organizational information processing perspective. Int. J. Prod. Econ.276:109369. doi: 10.1016/j.ijpe.2024.109369

  • CrossRef
  • Google Scholar

• 43

  YooY.BolandR. J.LyytinenK.MajchrzakA. (2012). Organizing for innovation in the digitized world. Organ. Sci.23, 1398–1408. doi: 10.1287/orsc.1120.0771

  • CrossRef
  • Google Scholar

• 44

  YooY.OlaH.LyytinenK. (2010). The new organizing logic of digital innovation: an agenda for information systems research. Inf. Syst. Res.21, 724–735.

  • Google Scholar

• 45

  ZhaoY.LiJ. (2026). Managerial climate risk perception bias and corporate investment inefficiency: evidence from Chinese listed firms. Finance Res. Lett.90:109399. doi: 10.1016/j.frl.2025.109399

  • CrossRef
  • Google Scholar

• 46

  ZhaoH.LiaoW. (2025). The nexus between climate risk and corporate green innovation: the moderating role of supply chain finance. Finance Res. Lett.85:107995. doi: 10.1016/j.frl.2025.107995

  • CrossRef
  • Google Scholar

• 47

  ZhouC.Richardson-BarlowC. (2026). A new framework for collaborative climate governance: linking consensus building and collective action. Policy Polit.54, 1–22. doi: 10.1332/03055736Y2026D000000089

  • CrossRef
  • Google Scholar