Introduction

Climate change has intensified the frequency and severity of extreme weather events globally, with extreme precipitation emerging as one of the most destructive meteorological phenomena. Recent studies indicate that these events will likely become more frequent and intense in the coming decades, causing significant economic losses exceeding thousands of billions of dollars annually and disrupting business operations, innovation activities, and social systems (Fu et al. 2023; Lesk et al. 2016; O’Gorman, 2015; Perera et al. 2020). Among various climate shocks, extreme precipitation warrants specific attention due to its distinctive characteristics and impact mechanisms on corporate activities. Unlike gradual climate changes or temperature anomalies, extreme precipitation events cause sudden, localized disruptions through multiple channels: direct physical damage to facilities and equipment, transportation network disruptions affecting supply chains, power outages compromising sensitive research processes, and employee mobility restrictions. These effects are particularly relevant for innovation activities, which often rely on specialized equipment, controlled environments, and consistent human capital inputs. While the macroeconomic impacts of extreme weather events have been extensively documented, their effects on firm-level innovation persistence remain understudied, despite innovation being crucial for maintaining competitive advantages and driving economic growth in an increasingly uncertain climate future (Lei and Xu 2025). The increasing frequency and intensity of extreme precipitation events make understanding their specific impacts on innovation persistence particularly urgent for both theoretical advancement and practical adaptation.

The urgency of understanding extreme precipitation’s impact on corporate innovation is underscored by several converging factors that collectively highlight the need for systematic investigation. Climate projections consistently indicate substantial increases in both the frequency and intensity of extreme precipitation events under various emission scenarios, suggesting that what we observe today may represent only the beginning of more severe disruptions to business operations (Wang et al. 2024; Zhou et al. 2019). Furthermore, while companies increasingly recognize the importance of sustained innovation for maintaining competitive advantage in rapidly evolving markets, their capacity to preserve consistent innovation efforts under environmental shocks appears to remain vulnerable and inadequately understood (Lei and Xu 2024). This knowledge gap becomes particularly concerning when considering that innovation persistence represents not merely a short-term operational challenge but a potentially crucial factor that may influence long-term organizational survival and growth in an increasingly uncertain climate future.

Current literature exhibits several significant gaps regarding the relationship between extreme precipitation and innovation persistence: First, while extensive research has documented the macroeconomic impacts of extreme weather events on agricultural production, economic growth, and financial markets (Chen and Gong 2021; Donadelli et al. 2021), these studies have not adequately addressed firm-level innovation responses. Second, although research on corporate innovation persistence has identified various determinants including technological intensity, market competition, and industry characteristics (Guarascio and Tamagni 2019; Hwang et al. 2023), the influence of environmental shocks remains largely unexplored. Third, while preliminary studies have examined immediate impacts of extreme weather on corporate operations through supply chain disruptions and physical asset damage (Benincasa et al. 2024; Wang et al. 2024), the specific mechanisms affecting long-term innovation persistence require systematic investigation. Fourth, the potential moderating roles of executive characteristics and geographic dispersion in maintaining innovation momentum under extreme weather conditions have not been adequately addressed in existing literature. Our focus on executive characteristics and geographic dispersion as moderating factors is theoretically grounded in organizational resilience literature. Executive characteristics, particularly R&D background and international experience, shape strategic decision-making under uncertainty, influencing how firms allocate resources during environmental disruptions. Geographic dispersion offers natural risk diversification, potentially insulating innovation activities from localized extreme weather events. These moderators represent actionable strategic levers that firms can adjust to enhance innovation resilience in an increasingly volatile climate.

To address these research gaps, our study aims to answer several key questions: (1) How does extreme precipitation affect corporate sustained innovation? (2) Through what potential channels is this impact transmitted? (3) How do executive characteristics and geographic dispersion moderate the relationship between extreme precipitation and corporate sustained innovation? By developing an extreme precipitation index using relative threshold methods and examining the mechanisms and moderating factors, this research sheds light on the interplay between natural disasters, corporate leadership, geographic dispersion, and innovation capabilities.

Our study makes several distinctive contributions to the literature by providing empirical evidence that advances our understanding of climate-innovation relationships in multiple dimensions. Our findings reveal that extreme precipitation significantly reduces both innovation input persistence and innovation output persistence, with the impact being more pronounced on innovation outputs, suggesting that environmental shocks not only disrupt resource allocation but also impair the efficiency of innovation processes themselves. Through comprehensive mechanism analysis, we demonstrate that extreme precipitation affects innovation persistence through two primary pathways: financial constraints and organizational resilience, with organizational resilience accounting for a larger proportion of the total effect than financial constraints. Our heterogeneity analysis uncovers that executive characteristics and geographic dispersion serve as crucial moderating factors, with executives possessing R&D backgrounds and international experience significantly mitigating the negative impacts, while geographic dispersion provides natural risk diversification that enhances innovation resilience. These empirical findings extend the disaster economics literature beyond immediate economic impacts to encompass long-term innovation capabilities, contribute to the innovation persistence literature by identifying environmental factors as important but previously unexplored influencing factors, and advance organizational resilience theory by demonstrating how executive human capital and geographic strategies interact with environmental shocks to influence innovation outcomes.

The remainder of this paper is organized as follows: Section “Literature Review and Hypothesis Development” reviews the relevant literature and develops our hypotheses. Section “Research Designs” describes our research design, including data collection, variable measurement, and empirical models. Section “Empirical Results” presents our empirical results and various robustness tests. Section “Potential Channel Analysis“ discusses the potential mechanism analysis. Finally, the paper is summarized, and recommendations are made accordingly.

Literature Review and Hypothesis Development

Literature review

As global climate change intensifies, the frequency and magnitude of extreme weather events, particularly extreme precipitation, have shown significant upward trends. Research indicates that these trends will likely intensify further in the coming decades, posing severe challenges to business operations and innovation activities (Fu et al. 2023; O’Gorman, 2015). Current studies have primarily investigated the macroeconomic impacts of extreme weather on agricultural production and economic growth, revealing that extreme precipitation affects economic activities not only through direct channels such as infrastructure damage and supply chain disruptions but also through indirect channels via financial market transmission mechanisms (Perera et al. 2020; Zhou et al. 2023). While these findings illuminate the broad impacts of extreme weather events on economic systems, the mechanisms through which they affect firm-level activities, particularly innovation, require further investigation.

As a core driver of economic growth, innovation derives its importance from its persistence in maintaining competitive advantages. Innovation persistence manifests not only in the stability of R&D investments but also in the accumulation of long-term innovation capabilities and the coherence of innovation strategies (Guarascio and Tamagni 2019). Prior research has demonstrated that innovation persistence is influenced by technological intensity, market competition, and industry characteristics (Hwang et al. 2023; Kaushik and Paul 2022). These studies have established the path-dependent nature of innovation persistence and its susceptibility to environmental changes. However, systematic theoretical analysis and empirical evidence regarding whether and how extreme precipitation affects firms’ ability to maintain consistent innovation input and output remain scarce.

Existing literature has examined the impact of extreme weather on corporate innovation from several perspectives. From a resource allocation perspective, firms must divert resources to post-disaster reconstruction and production recovery, thereby crowding out innovation investments (Lei and Xu 2024). From a human capital perspective, extreme precipitation may lead to R&D personnel turnover or reduced work efficiency, affecting innovation activities (Bai et al. 2023). From a supply chain perspective, extreme weather can disrupt upstream and downstream cooperation, destabilizing innovation networks (Wang et al. 2024). While these studies have revealed certain channels through which extreme weather affects corporate innovation, two critical issues remain unaddressed: First, existing research has primarily focused on changes in absolute levels of innovation input and output, with insufficient attention to innovation persistence as a crucial indicator of long-term innovation capability; second, the internal mechanisms through which extreme precipitation affects innovation persistence, particularly the potential key transmission channels of financial constraints and organizational resilience, remain unexplored.

Regarding financial constraints, while studies have found that extreme weather affects corporate financing behaviour (Benincasa et al. 2024), how this impacts firms’ ability to maintain persistent innovation investment remains unclear. Theoretically, extreme precipitation may increase firms’ operational risks and financing costs, forcing them to increase liquidity reserves and thereby affecting the stability of long-term innovation investments. While existing research identifies organizational resilience as a crucial capability for firms to cope with external shocks (Li et al. 2024), its role in mediating the impact of extreme weather on innovation persistence requires further investigation. Organizational resilience may moderate the impact of extreme precipitation on innovation persistence through its influence on strategic determination, resource integration capabilities, and risk management effectiveness.

Regarding firm characteristics, preliminary research has found that geographic diversification helps firms disperse operational risks (Du et al. 2024), while managerial characteristics influence corporate innovation decisions (Akcigit et al. 2022). However, how these characteristics affect firms’ ability to maintain innovation persistence under extreme weather shocks remains unverified. Specifically, geographic dispersion, as a crucial strategic characteristic, may influence firms’ innovation resilience through resource allocation optimization, risk diversification, and enhanced shock resistance. Similarly, executives’ R&D background and international experience, as key human capital elements, may moderate firms’ ability to cope with extreme weather shocks through their influence on strategic decision-making and risk management effectiveness. These potential moderating effects warrant further investigation.

Based on the above analysis, this study extends the existing literature by proposing and systematically examining how extreme precipitation affects corporate innovation persistence through the dual mechanisms of financial constraints and organizational resilience. Specifically, we posit that the financial constraint mechanism manifests primarily through extreme precipitation’s impact on operational risks and financing costs, which alters firms’ cash holdings and investment decisions, thereby affecting innovation persistence. The organizational resilience mechanism reflects firms’ ability to maintain innovation stability through organizational structure adjustment, resource reconfiguration, and management innovation when facing extreme weather shocks. This theoretical framework helps understand the underlying causes of compromised innovation persistence and provides new insights for firms to develop preventive measures.

Furthermore, this study examines the moderating effects of geographic dispersion and executive characteristics. We argue that firms with higher geographic dispersion can better maintain innovation persistence by allocating resources across regions to diversify risks and reduce dependence on single-region extreme weather events. Executives with R&D backgrounds and international experience may leverage their professional knowledge and global perspective to make more informed strategic decisions under extreme weather shocks, helping firms maintain stable innovation investments. These findings provide important implications for firms to enhance innovation resilience by optimizing the geographic layout and talent configuration in the context of climate change.

Regarding theoretical contributions, this study enriches research on determinants of corporate innovation persistence while extending the microeconomic investigation of extreme weather consequences. By constructing a dual-mechanism framework of financial constraints and organizational resilience, we deepen the understanding of how extreme weather affects corporate innovation activities. In terms of practical implications, our findings provide important references for firms to formulate innovation strategies and risk management measures, particularly in optimizing the geographic layout and management team configuration, which will help enhance firms’ ability to cope with climate change.

Hypotheses development

Building upon the preceding analysis, this study develops a theoretical framework and corresponding hypotheses from organizational economics and strategic management theory perspectives. According to the resource-based view, a firm’s competitive advantage stems from its heterogeneous resources, with sustained innovation capability as a unique strategic resource, playing a decisive role in maintaining competitive advantage (Barney 2000). The formation and maintenance of this capability require firms to make long-term stable investments in R&D funding, talent teams, and organizational management. However, in the context of global climate change, extreme precipitation, as a significant external shock, may affect firms’ ability to maintain innovation investments through multiple mechanisms.

Extreme precipitation and innovation persistence

According to the Trade-off Theory, firms must balance different objectives in their resource allocation process. When facing operational pressures from extreme precipitation, firms may be forced to adjust their resource allocation strategies, redirecting resources originally intended for innovation activities toward post-disaster reconstruction and production recovery (Lei and Xu 2024). This adjustment in resource allocation priorities directly affects current R&D investments and may disrupt the continuity of firms’ innovation activities, impacting the accumulation of long-term innovation capabilities. From an organizational learning perspective, developing innovation capabilities requires continuous learning and practice. Once innovation activities are interrupted, this can create gaps in knowledge accumulation and potentially lead to the dissolution of innovation teams and the breakdown of innovation networks. These effects require considerable time to recover from (Guarascio and Tamagni 2019).

Furthermore, based on behavioural theory, corporate managers exhibit risk-averse tendencies when facing external shocks. Extreme precipitation increases uncertainty in innovation activities, potentially prompting managers to adopt more conservative investment strategies and reduce investments in high-risk innovation projects (Wang et al. 2024). This strategic shift affects the progression of individual innovation projects and may also lead to adjustments in overall corporate innovation strategy, impacting the systematicity and continuity of innovation activities. In practice, extreme precipitation often leads firms to postpone or cancel planned R&D projects and reduce innovation investments, which are detrimental to maintaining innovation persistence. Based on the above analysis, we propose:

H1: Extreme precipitation significantly reduces corporate innovation persistence.

Financial constraint mechanism

According to information asymmetry theory, external shocks exacerbate the information asymmetry between firms and external investors. During extreme precipitation events, external investors struggle to accurately assess the extent of corporate damage and recovery capability, leading them to demand higher risk compensation, thereby increasing firms’ financing costs (Benincasa et al. 2024). Additionally, according to precautionary motive theory, firms tend to hold more cash to address potential risks when facing increased uncertainty. Research has found that firms experiencing extreme weather shocks often increase their cash holdings, and this conservative financial strategy further reduces resources available for innovation investments (Huang et al. 2018).

From a corporate life cycle theory perspective, innovation activities require sustained and stable financial input to generate expected outcomes. However, intensified financial constraints may force firms to reduce or interrupt R&D investments, affecting the continuity of innovation projects. This is particularly true for important but long-term basic research projects, which are more likely to be suspended or cancelled under increased financing constraints. Such disruptions not only affect current innovation output but may also damage firms’ long-term innovation capabilities (Wang et al. 2024). Moreover, financial constraints may lead to the loss of core R&D talent through their impact on compensation levels and benefits, further affecting the sustained development of innovation activities. Based on these arguments, we propose:

H2: Financial constraints mediate the relationship between extreme precipitation and corporate innovation persistence.

Organizational resilience mechanism

Dynamic capability theory suggests that firms must continuously integrate, build, and reconfigure internal and external resources to address environmental changes. As a crucial dynamic capability, organizational resilience reflects a firm’s ability to maintain normal operations and strategic objectives when facing external shocks (Teece 2018). However, extreme precipitation, as a sudden external shock, may weaken firms’ organizational resilience through multiple channels.

First, from an operational perspective, extreme precipitation often results in damaged infrastructure, disrupted production processes, and impaired supply chain coordination. These sudden impacts directly affect firms’ daily operations and may cause organizational structure disorder and inter-departmental coordination difficulties (Lei and Xu 2024). Particularly when firms lack experience and contingency plans for extreme weather, organizational and decision-making efficiency may significantly decline, weakening their ability to maintain normal operations. Second, from a resource integration perspective, extreme precipitation increases uncertainty in corporate resource allocation. Firms must redirect resources originally intended for long-term strategic activities such as innovation toward post-disaster reconstruction and production recovery, and this passive resource reallocation weakens firms’ strategic execution capability (Wang et al. 2024).

Meanwhile, extreme weather may damage firms’ connections with external partners, affecting their ability to acquire and integrate external resources and reducing organizational resilience. Third, from an employee morale and organizational climate perspective, frequent extreme weather events may affect employee work enthusiasm and organizational identification. Research has found that deterioration in working conditions and increased safety risks due to extreme weather reduce employee work efficiency and organizational commitment (Bai et al. 2023). These negative impacts may damage corporate organizational culture and team cohesion, weakening organizational response capabilities.

When organizational resilience is weakened, firms’ ability to maintain innovation activities is also affected. Firms with weaker organizational resilience often demonstrate poor strategic determination and are more likely to alter established innovation development directions due to short-term pressures. Additionally, due to the lack of effective risk management systems and emergency response mechanisms, these firms struggle to protect innovation activities from extreme weather impacts, leading to fluctuations and interruptions in innovation investments (Li et al. 2024). Conversely, firms that maintain strong organizational resilience can sustain innovation activities even under extreme weather conditions. Accordingly, we propose:

H3: Organizational resilience mediates the relationship between extreme precipitation and corporate innovation persistence.

The moderating effect of geographic dispersion

According to risk diversification theory, geographic diversification is an important strategy for firms to disperse operational risks. Geographically dispersed firms can reduce their dependence on extreme weather in any region through cross-regional resource allocation, enhancing overall risk resistance capability (Du et al. 2024). From a resource dependence theory perspective, geographic diversification also helps firms access diverse innovation resources and establish broader external cooperation networks, all contributing to maintaining innovation stability.

Geographic dispersion may moderate the impact of extreme precipitation on corporate innovation persistence through several aspects: First, at the resource allocation level, geographic diversification enables firms to flexibly allocate cross-regional resources during extreme weather events, such as temporarily transferring R&D tasks from affected areas to other regions, thereby reducing the risk of innovation activity interruption. Empirical research has found that geographically dispersed firms demonstrate stronger operational resilience when facing natural disasters (Du et al. 2024). Second, at the risk prevention level, geographic diversification provides firms with a natural “insurance” mechanism. Due to differences in the timing and severity of extreme weather across regions, firms with high geographic dispersion can stabilize overall innovation investment through regional complementarity and balance. This diversification strategy reduces the impact of extreme weather in any single region and maintains the continuity of overall corporate innovation activities. Research indicates that firms with higher geographic diversification demonstrate stronger financial stability under external shocks (Huang et al. 2018). Third, from an innovation network perspective, geographic dispersion allows firms to access innovation resources and partners across different regions. This diversified innovation network enhances firms’ innovation capabilities and strengthens their risk-resistance capacity in innovation activities. Even if innovation networks in one region are affected by extreme weather, firms can maintain innovation activities through networks in other regions. Moreover, innovation resources and experiences from different regions can be cross-referenced, helping firms better respond to challenges posed by extreme weather. Based on this analysis, we propose:

H4: Geographic dispersion mitigates the negative impact of extreme precipitation events.

The moderating effect of executive characteristics

Upper echelons theory suggests that firms’ strategic choices and organizational outcomes largely reflect their top management teams’ cognition and value orientations. Managers’ educational backgrounds, work experiences, and other characteristics influence their perception of the external environment and strategic decision-making (Carpenter et al. 2004). In this study, executives’ R&D backgrounds and international experience may moderate the impact of extreme precipitation on innovation persistence through the following mechanisms:

First, from a cognitive perspective, executives with R&D backgrounds typically have a deeper understanding of innovation activities and a clearer recognition of innovation investments’ long-term value and interruption risks. This professional cognition makes them more inclined to maintain stable innovation investments rather than excessively cutting R&D expenditure when facing extreme weather shocks. Empirical research has also found that executives with technical backgrounds demonstrate stronger forward-looking capabilities in formulating and implementing innovation strategies (Akcigit et al. 2022). Second, from a capability perspective, executives with international experience often possess richer risk management experience and broader resource acquisition channels. This experience and these resources help them make more scientific response decisions under extreme weather shocks, such as seeking alternative R&D resources through international cooperation networks or drawing on international best practices to improve risk prevention mechanisms. Research shows that executives’ international experience can significantly enhance firms’ risk management capabilities and innovation performance (Lei and Xu, 2024). Third, from an organizational culture perspective, executives with R&D backgrounds and international experience are more likely to foster innovation-oriented organizational cultures and establish institutional systems supporting sustained innovation within their firms. Such organizational cultures and institutional arrangements can provide a “buffer” for innovation activities under extreme weather shocks, helping firms maintain stable innovation investments. These executives typically demonstrate stronger strategic determination and are less likely to alter innovation development strategies due to short-term difficulties. Accordingly, we propose:

H5: The negative impact of extreme precipitation events is attenuated by executives’ R&D background and international experience.

Finally, the whole framework of the research hypotheses is shown in Fig. 1, and later on the empirical tests go to verify whether the hypotheses are valid or not.

Fig. 1: Research Framework.
figure 1

This figure illustrates the theoretical framework showing the relationships between extreme precipitation, innovation persistence, mediating mechanisms (financial constraints and organizational resilience), and moderating factors (executive characteristics and geographic dispersion).

Full size image

Research Design

Key variables

Our examination of innovation persistence centres on two complementary dimensions: innovation input and output continuity. Following established methodologies in the Chinese innovation literature (He et al. 2017; Zhao and Qi 2023), we construct measures that capture both the relative dynamics and absolute scale of innovation activities. While conventional approaches might use geometric averages of growth rates, our measures deliberately incorporate magnitude to reflect the substantive importance of innovation activities in the Chinese corporate context.

The first dependent variable, innovation input persistence (\({{\rm{IIP}}}_{\left(i,t\right)}\)), captures the stability and growth of firms’ R&D investments. Specifically, \({{\rm{IIP}}}_{\left(i,t\right)}\) is measured as the growth rate of R&D expenditures in years t and t-1 relative to the sum in years t-2 and t-1, multiplied by the total R&D expenditures in years t and t-1. To address the dimensional nature of this measure and ensure comparability across firms of different sizes, we employ a logarithmic transformation in our empirical specifications.

$${{\rm{IIP}}}_{\left(i,t\right)}=\left(\frac{{{\rm{IIN}}}_{\left(i,t\right)}+{{\rm{IIN}}}_{\left(i,t-1\right)}}{{{\rm{IIN}}}_{\left(i,t-1\right)}+{{\rm{IIN}}}_{\left(i,t-2\right)}}\right)\times \left({{\rm{IIN}}}_{\left(i,t\right)}+{{\rm{IIN}}}_{\left(i,t-1\right)}\right)$$
(1)

Building on this input-focused measure, our second dependent variable, innovation output persistence (\({{\rm{OIP}}}_{\left(i,t\right)}\)), extends the analysis to capture the consistency of innovation outcomes through patent applications. Following the same methodological approach, \({{\rm{OIP}}}_{\left(i,t\right)}\) is also log-transformed in our empirical analysis to ensure proper scaling and cross-firm comparability.

$${{\rm{OIP}}}_{\left(i,t\right)}=\left(\frac{{{\rm{OIN}}}_{\left(i,t\right)}+{{\rm{OIN}}}_{\left(i,t-1\right)}}{\left({{\rm{OIN}}}_{\left(i,t-1\right)}+{{\rm{OIN}}}_{\left(i,t-2\right)}\right)}\right)\times \left({{\rm{OIN}}}_{\left(i,t\right)}+{{\rm{OIN}}}_{\left(i,t-1\right)}\right)$$
(2)

Having established these measures of sustained innovation activity as our dependent variables, we turn to quantifying the environmental shocks that may disrupt them. The cornerstone of our analysis lies in developing a robust extreme precipitation index that can effectively capture the intensity and frequency of extreme precipitation events while accounting for regional climate variations. Following international meteorological standards, particularly the guidelines established by the World Meteorological Organization (WMO) and the World Climate Research Programme (WCRP), we employ a relative threshold method to measure extreme precipitation events. This approach, supported by recent literature (Du et al. 2014; Ma et al. 2024), enables us to control for natural climatic differences across regions while maintaining consistency in measurement.

Our methodology draws specifically from the “very heavy precipitation” indicator developed by the Expert Team on Climate Change Detection and Indices (ETCCDI), which has gained widespread acceptance in climate research. This indicator quantifies extreme precipitation by measuring the total annual precipitation that occurs on days when daily precipitation exceeds the 99th percentile of the precipitation distribution during a predefined baseline period. In our study, we establish the baseline period as 1982–2011, providing a robust 30-year climatological reference against which we assess extreme events in our sample period of 2012–2022.

The mathematical formulation of our extreme precipitation index (Precipitation) is expressed as:

$$\begin{array}{c}{{Precipitation}}_{i,t}=\mathop{\sum }\limits_{d=1}^{n}\left({{precipitation}}_{i,t,d}-{P}_{99W30}{{Precipitation}}_{i}\right),\\ {{{Where}}\; {{rainfall}}}_{i,t,d} > {{P}_{99W30}{{Rainfall}}}_{i}\end{array}$$
(3)

In this specification, \({{Precipitation}}_{i,t}\) represents the extreme precipitation intensity for the city where firm i is located in year t, while \({{precipitation}}_{i,t,d}\) denotes the daily precipitation for that city on day d of year t. The threshold \({P}_{99W30}{{Precipitation}}_{i}\) represents the 99th percentile of daily precipitation calculated over the 30-year baseline period (1982–2011) for the respective city. This local threshold approach ensures that our measure of extreme precipitation is contextually appropriate for each region’s typical climate patterns.

The construction process aggregates precipitation that exceeds the local threshold throughout the year, effectively capturing both the intensity of individual extreme precipitation events and their cumulative impact over time. To enhance the interpretability of our results, we scale the final values by a factor of 100, which provides a more intuitive measure of relative differences in extreme precipitation intensity across different cities and years without affecting the underlying relationships in our analysis.

This methodological approach offers several advantages. First, using a relative threshold based on local climate conditions, we consider that what constitutes “extreme” precipitation varies significantly across China’s diverse climate regions. Second, the long baseline period ensures that our threshold values are stable and represent long-term climate patterns. Third, by considering the frequency and intensity of extreme events, our index provides a more comprehensive measure of extreme precipitation exposure than simpler metrics based solely on occurrence counts or absolute thresholds.

Data sources

Our empirical analysis draws upon a comprehensive dataset combining information from three authoritative sources: the China Stock Market & Accounting Research (CSMAR) database, which provides detailed firm-level financial and operational data; the China National Meteorological Science Data Center (CNMDC), which offers extensive meteorological observations; and the National Intellectual Property Administration (NIPA), which provides patent application and grant information. The final sample comprises 19,165 firm-year observations from Chinese public companies from 2012 to 2022, following the exclusion of observations with missing or anomalous values.

The matching between meteorological data and firm locations was conducted with particular attention to spatial accuracy. For each firm, geographic coordinates were obtained based on the registered headquarters address and verified against company filings. These coordinates were then used to extract precipitation data from the gridded dataset at the precise location. For multi-region firms, we primarily used headquarters locations, as strategic innovation decisions typically emanate from central management. However, we acknowledge this as a potential limitation and address it through our geographic dispersion analysis. The data cleaning process followed systematic procedures to ensure sample integrity. Observations were excluded if they contained missing values for core variables (innovation measures, precipitation index, or key control variables) or exhibited extreme outliers defined as values exceeding third standard deviations from the mean after winsorization at the 1% and 99% levels. This approach balanced the need to remove potentially erroneous data points while preserving the representativeness of the sample.

Modeling

The processing of meteorological data requires special attention due to its spatial characteristics. The raw weather data consists of daily observations from weather stations across China. Following established methods in climate science literature and drawing from the research results of local Chinese scholars (Liu et al. 2021), the Inverse Distance Weighting (IDW) interpolation method was employed to convert station-level observations into a gridded dataset. This interpolation process first transforms point-based weather station data into continuous meteorological variable surfaces, which are then aggregated to the prefecture-level using administrative boundaries. While alternative methods such as ordinary kriging offer theoretical advantages in certain contexts, IDW provides superior performance for precipitation data in our study area due to its ability to preserve local variation patterns and its computational efficiency when dealing with large datasets. The baseline model is specified as follows:

Our baseline econometric specification is formulated as follows:

$$\begin{array}{ll}{{Innovation}}_{i,t}={\beta }_{0}+{\beta }_{1}{\text{Precipitation}}_{i,t}\\\qquad\qquad\qquad+\gamma {X}_{i,t}+{\theta {\rm{Climate}}}_{i,t}+{\alpha }_{i}+{\delta }_{t}+{\epsilon }_{i,t}\end{array}$$
(4)

Where \({{Innovation}}_{i,t}\) represents either the innovation input persistence (IIP) or innovation output persistence (OIP) for firm i in year t, and \({\text{Precipitation}}_{i,t}\) denotes our constructed extreme precipitation index. The model incorporates extensive control variables at the firm and city levels to account for potential confounding factors. Firm-level controls include size (measured as the natural logarithm of total assets), leverage (total debt divided by total assets), return on assets (ROA), firm age (natural logarithm of the number of years since incorporation), and ownership structure (percentage of state ownership). City-level controls encompass GDP per capita, population density, and climate variables. The Climate variables represent a comprehensive set of climate-related control variables that could affect innovation persistence. These include annual mean temperature, annual temperature standard deviation, and annual average relative humidity measured at the city level where firm i is located in year t. The inclusion of these annual climate controls is crucial as they help isolate the specific effect of extreme precipitation from other weather-related phenomena that might influence corporate innovation activities. Additionally, controlling for these climate variables addresses potential concerns about omitted variable bias, as extreme precipitation events may be systematically correlated with other annual weather patterns. To address potential endogeneity concerns and control for unobserved heterogeneity, our specification includes firm fixed effects (\({\alpha }_{i}\)) and year fixed effects (\({\delta }_{t}\)). The firm-fixed effects capture time-invariant firm characteristics that might influence innovation persistence, while year-fixed effects control for macroeconomic conditions and other temporal factors affecting all firms simultaneously. The error term \({\epsilon }_{i,t}\) is clustered at the city level to account for potential spatial correlation in the residuals.

Our choice of firm and year fixed effects is theoretically and empirically justified by the need to control for time-invariant firm characteristics and time-varying macroeconomic factors. The firm fixed effects control for unobserved heterogeneity at the firm level, including industry-specific characteristics, organizational culture, and innovation capabilities that remain relatively stable over time. Year fixed effects account for economy-wide shocks such as policy changes, technological breakthroughs, and macroeconomic conditions that affect all firms simultaneously. The Hausman test results (χ² = 187.34, p < 0.01) further confirm that fixed effects estimation is more appropriate than random effects for our analysis, as the null hypothesis of no systematic difference between the two models is rejected. Standard errors are clustered at the city level to account for spatial correlation in the residuals, as firms within the same geographic area may be subject to common unobserved shocks. This approach addresses potential concerns about within-group error correlation and produces more conservative inference.

Empirical Results

Descriptive statistical analysis

Table 1 presents descriptive statistics for the main variables. During the study period, the means of innovation input persistence (IIP) and innovation output persistence (OIP) are 0.486 and 0.512, with standard deviations of 0.352 and 0.378, respectively, reflecting significant variations in innovation activity maintenance among Chinese listed companies. Innovation input persistence ranges from 0 to 1.892, indicating that while some firms completely suspended innovation inputs during the study period, others maintained high input growth. The innovation output persistence indicator shows similar distributional characteristics, ranging from 0 to 2.015, suggesting marked divergence in firms’ innovation output continuity.

Table 1 Descriptive Statistics.
Full size table

The sample firms have a mean size (Size) of 22.145, ranging from 19.234 to 26.789, indicating coverage of listed companies of various scales. The mean leverage ratio (Leverage) of 0.432 suggests moderate debt levels among sample firms. The profitability indicator ROA has a mean of 0.056, ranging from −0.234 to 0.321, reflecting significant variations in firms’ profitability. The distributional characteristics of firm age (Age) and state ownership percentage (State) indicate that the sample encompasses firms at different developmental stages and ownership types. Regarding environmental indicators, the extreme precipitation index (Rainfall) has a mean of 2.345, standard deviation of 1.967, and maximum value of 9.234. These distributional characteristics confirm the occurrence of multiple significant heavy precipitation events during the study period, with substantial variations in extreme precipitation intensity across regions and years. City-level control variables show sufficient variation in economic development (GDP), population density (Population), and climate conditions (Temperature, Temp std, and Humidity) across the sample firms’ locations, which helps control for regional differences in the research findings.

Benchmark model results

As shown in Table 2, our empirical analysis presents compelling evidence of how extreme precipitation events significantly impact corporate sustained innovation through various channels. The third and fourth columns respectively added potential influencing factors on the basis of the first and second columns. We found that the core relationship was still statistically significant, strengthening the robustness of our research results. The baseline regression results demonstrate a substantial negative relationship between extreme precipitation and innovation persistence, with the coefficients of Precipitation being −0.1652 (p < 0.01) for innovation input persistence (IIP) and −0.1923 (p < 0.05) for innovation output persistence (OIP). These statistically significant findings suggest that a one-unit increase in the precipitation index leads to considerable reductions of 16.52% in innovation input persistence and 19.23% in innovation output persistence.

Table 2 The Impact of Extreme Precipitation on Corporate Continuous Innovation.
Full size table

The differential magnitude between input and output persistence reveals an important nuance in how environmental shocks affect corporate innovation. The larger impact on output persistence (−0.1923 versus −0.1652) suggests that extreme precipitation events not only disrupt resource allocation but also impair the efficiency of the innovation process itself. For instance, during the severe flooding in the Yangtze River Delta region in 2020, high-tech firms experienced cascading disruptions where semiconductor manufacturers maintained partial R&D funding but faced disproportionate impacts on their ability to conduct precise experiments due to equipment sensitivity and clean room environmental requirements. Similarly, biotechnology firms found that even when maintaining research budgets, their experimental processes were significantly disrupted by environmental instability, leading to lower innovation productivity.

To address potential multicollinearity concerns, we calculated variance inflation factors (VIFs) for all variables in our model. The VIF values for our precipitation index and climate-related variables range from 1.24 to 2.87, well below the conventional threshold of 10, indicating that multicollinearity does not significantly affect our estimates. The correlation between the precipitation index and humidity is 0.32, suggesting that these variables capture distinct climate dimensions.

Robustness checks

To ensure the validity and reliability of our findings, we conducted a comprehensive series of robustness tests. As shown in Table 3, when employing alternative measures of extreme precipitation, the 95th percentile threshold (Precipitation95) yields coefficients of −0.0754 (p < 0.05) for IIP and −0.0846 (p < 0.10) for OIP. In contrast, the maximum 5-day precipitation measure (Max5days) shows stronger effects of −0.2021 and −0.2453, respectively (both at p < 0.01). These progressively larger coefficients with more extreme measures reveal an important non-linear relationship between weather severity and innovation disruption. This pattern was particularly evident during the 2021 Henan floods, where technology parks experienced exponentially greater disruptions as precipitation intensity increased beyond critical thresholds.

Table 3 Replacing Core Explanatory Variable and Lag Terms.
Full size table

While we recognize that extreme precipitation effects may manifest with some delay, our primary specification focuses on contemporaneous effects for several reasons. First, innovation decisions typically involve real-time resource allocation adjustments in response to environmental shocks. Second, financial markets rapidly incorporate information about extreme weather events, affecting firms’ financing conditions immediately. Nevertheless, we conducted additional tests using Lag one-stage Precipitation measures (LAGPRE) and found consistent results (coefficient for IIP: −0.1423, p < 0.05; for OIP: −0.1612, p < 0.05), though with slightly smaller magnitudes, suggesting that immediate impacts are more pronounced than delayed effects.

To address potential endogeneity concerns, we implement an instrumental variable approach using the average city slope as an instrument for extreme precipitation. This choice is grounded in fundamental principles of hydrogeology, as sloped terrain forces moist air masses upward (orographic lifting), leading to enhanced precipitation through adiabatic cooling and condensation. The instrument satisfies both the relevance condition, confirmed by our high F-statistic of 153.62 (substantially exceeding the conventional weak instrument threshold of 10) and the exclusion restriction, as the average slope is determined by geological processes over millions of years and is demonstrably exogenous to current economic activities. The IV estimation results, presented in Table 4, corroborate our main findings and pass the weak instrument test (F-statistic: 153.62), further substantiating the robustness of our conclusions. A potential concern with our instrumental variable approach is that the time-invariant nature of average city slope might be absorbed by firm fixed effects. This concern is mitigated by our specific empirical strategy. While slope itself is time-invariant, our instrument operates through the interaction between slope and precipitation patterns, which vary substantially over time. In the first-stage regression, we estimate how slope affects the temporally varying occurrence of extreme precipitation, allowing for identification even in the presence of firm fixed effects. The strong F-statistic of 153.62 empirically confirms that this time-varying relationship provides sufficient identifying variation. Furthermore, our approach is econometrically valid because the endogenous variable (extreme precipitation) shows substantial within-firm variation over time, allowing for identification of its effects even when instrumenting it with a time-invariant geographical feature.

Table 4 Instrumental Variables Test Results.
Full size table

To further validate the robustness of our findings, we examine the temporal dynamics of how extreme precipitation affects corporate innovation persistence. Using an event study approach, we designate the year prior to a firm’s first major extreme precipitation event (defined as precipitation index exceeding the sample period median) as the baseline period. As shown in Fig. 2, at the 95% confidence interval, the estimated coefficients for periods preceding major extreme precipitation events are statistically insignificant, indicating no systematic anticipatory effects. However, following these events, the coefficients become significantly negative and exhibit a persistent downward trend, corroborating our baseline findings. These results provide strong support for the causal relationship between extreme precipitation and corporate innovation persistence.

Fig. 2: Event study analysis.
figure 2

This figure presents the estimated coefficients for periods before and after major extreme precipitation events at 95% confidence interval, demonstrating the temporal dynamics of extreme precipitation’s impact on corporate innovation persistence.

Full size image

A placebo test was conducted to mitigate further concerns of omitted variable bias. By randomly reassigning extreme precipitation data across our sample, we demonstrate that these randomized assignments yield coefficients not significantly different from zero, reinforcing the robustness of our baseline regression results and confirming that spurious correlations or unobserved factors do not drive our findings. As shown in Fig. 3, the distribution of placebo test coefficients is centered around zero and exhibits a normal distribution pattern, while our actual estimated coefficient (indicated by the black line) lies well outside this null distribution. This stark contrast between the actual and placebo coefficients provides strong evidence.

Fig. 3: Placebo test.
figure 3

This figure displays the distribution of placebo test coefficients centered around zero with a normal distribution pattern, with the actual estimated coefficient (indicated by the black line) lying outside this null distribution, confirming that our findings are not driven by spurious correlations.

Full size image

To further validate our findings, we employed a difference-in-differences (DID) framework focusing on a major precipitation shock—the 2021 Henan floods. This approach allows us to more precisely identify the immediate and lagged effects of a specific extreme precipitation event. We classified firms located in Henan province as the treatment group and constructed a control group using propensity score matching based on pre-flood control variables characteristics. The DID specification is:

$$\begin{array}{ll}{{Innovation}}_{i,t}={\beta }_{0}+{\beta }_{1}{{Treatment}}_{i}+{\beta }_{2}{{Post}}_{t}+{\beta }_{3}\left({{Treatment}}_{i}\times {{Post}}_{t}\right)\\\qquad\qquad\qquad+\gamma {X}_{i,t}+{\alpha }_{i}+{\delta }_{t}+{\epsilon }_{i,t}\end{array}$$
(5)

Where Treatment is a dummy variable equal to 1 for firms in Henan province and 0 otherwise, and Post equals 1 for periods after the 2021 floods and 0 otherwise. The coefficient of interest, \({\beta }_{3}\), captures the differential impact of the floods on treated firms’ innovation persistence.

The significant negative coefficients for the interaction term in Table 5 confirm that firms exposed to the Henan floods experienced substantially larger decreases in innovation persistence compared to matched firms in unaffected regions. This result provides additional causal evidence supporting our main findings.

Table 5 Analysis of 2021 Henan Floods.
Full size table

Heterogeneity analysis

Further heterogeneity analysis reveals the significant moderating role of firm characteristics in the impact of extreme precipitation. Regarding executive characteristics, R&D background (RDB) is a dummy variable that equals 1 if executives have education or work experience in R&D-related fields, and 0 otherwise. Similarly, international experience (IE) is a dummy variable that equals 1 if executives have overseas education or work experience, and 0 otherwise. Geographic dispersion (GD) is measured following previous research (García and Norli 2012; Shi et al. 2015), constructed using the geographic distribution information of firms’ subsidiaries and affiliates (including subsidiaries, sub-subsidiaries, associated companies, and joint ventures) from Wind database’s “Multi-dimensional Data” and “In-depth Information”. Specifically, we identify the locations of these companies and verify them against company websites and annual reports, calculating the distribution across 31 provinces. The geographic dispersion indicator is the natural logarithm of the number of provinces plus one. GD equals 1 if a firm’s geographic dispersion is above the sample median, and 0 otherwise.

Table 6 demonstrates that executive characteristics significantly influence firms’ ability to cope with environmental shocks. Specifically, executives with R&D backgrounds better maintain firms’ innovation momentum, as evidenced by the significantly positive coefficient of the interaction between precipitation and R&D background (RDB). This finding suggests that professional backgrounds enable executives to more accurately assess the strategic value of innovation projects and make more targeted trade-offs when resources are constrained. Similarly, executives with international experience (IE) demonstrate stronger risk management capabilities, reflected in the significantly positive interaction effect coefficient. This may stem from their diverse risk management experience accumulated through international operations.

Table 6 The Moderating Role of Executive Characteristics.
Full size table

Table 7 further indicates that geographic distribution strategies play a crucial role in firms’ response to environmental shocks. The significantly positive coefficient of the interaction between geographic dispersion and extreme precipitation suggests that firms with higher geographic concentration face greater innovation disruption risks. This finding emphasizes the importance of geographic diversification in mitigating environmental risks and provides crucial implications for firms’ spatial layout strategies. Highly geographically concentrated firms are more vulnerable to localized extreme weather events, while geographic dispersion provides a natural risk hedging mechanism. To further explore the heterogeneous effects of extreme precipitation, we examined differences between state-owned enterprises (SOEs) and non-SOEs. As shown in Table 7, the negative impact of extreme precipitation on innovation persistence is significantly stronger for non-SOEs compared to SOEs, with interaction coefficients of 0.0724 (p < 0.01) for IIP and 0.0812 (p < 0.01) for OIP. This heterogeneity likely stems from SOEs’ superior access to government resources and implicit guarantees, which provide a buffer against environmental shocks. These findings suggest that ownership structure plays a crucial role in determining firms’ vulnerability to climate events and their ability to maintain innovation momentum.

Table 7 The Moderating Role of Geographical Dispersion.
Full size table

Potential Channel Analysis

Analysis of the internal mechanisms through which extreme precipitation affects innovation persistence requires examining specific transmission pathways. Following the previous research (Lei and Xu 2024), we measure financing constraints (FC) using the KZ index, where higher values indicate greater financing constraints. Organizational resilience (OR) was measured according to the methodology studied by local Chinese scholars (Wu and Feng 2022), incorporating performance growth and volatility dimensions. Performance growth is measured by cumulative sales revenue growth over the past three years, while volatility is measured by the standard deviation of monthly stock returns within one year. The composite organizational resilience indicator is constructed by standardizing these dimensions and subtracting volatility from growth, with higher values indicating stronger organizational resilience.

Table 8 reports the transmission path analysis through financing constraints. Column (1) shows that extreme precipitation events significantly increase firms’ financing constraints (coefficient: 0.3818). This finding indicates that extreme precipitation intensifies financial pressure by damaging physical assets and disrupting normal operations. Columns (2) and (3) further reveal that heightened financing constraints significantly reduce innovation persistence, with inhibitory effects of −0.0220 and −0.0141 on innovation input and output, respectively. This suggests that firms facing severe financing constraints experience significantly diminished ability to maintain stable innovation investments, often manifesting as reduced R&D spending or delayed innovation projects to address short-term liquidity pressures.

Table 8 Financing Constraint Transmission Path.
Full size table

Table 9 demonstrates the organizational resilience transmission pathway. Column (1) shows that extreme precipitation significantly weakens organizational resilience (coefficient: −0.2758), reflected in reduced sales growth momentum and increased stock return volatility. Columns (2) and (3) indicate positive effects of organizational resilience on innovation input and output persistence of 0.0638 and 0.0598, respectively, suggesting that more resilient organizations better maintain innovation stability during extreme precipitation events.

Table 9 Organizational Resilience Transmission Path.
Full size table

To formally test the mediation effects, we employed the Karlson-Holm-Breen (KHB) method, which is specifically designed to decompose total effects into direct and indirect effects. The KHB analysis confirms that financing constraints mediate 5.08% of the total effect of extreme precipitation on innovation input persistence and 2.81% of the effect on output persistence. Similarly, organizational resilience accounts for 10.65% and 8.58% of the total effects on input and output persistence, respectively. These results provide statistical validation for our proposed dual-mechanism framework.

These findings provide crucial insights into the transmission mechanisms of extreme precipitation’s impact on corporate innovation. The dual pathways of financing constraints and organizational resilience indicate that the impact on innovation activities is a complex process involving both financial resources and organizational capabilities. This understanding has important implications for corporate strategy development: firms need to establish adequate financial buffers while enhancing organizational resilience through improved sales growth capabilities and reduced operational volatility. As climate change intensifies, cultivating these response capabilities becomes increasingly crucial for sustainable corporate development.

Conclusions and Implications

Result discussion

This study provides comprehensive evidence on how extreme precipitation affects corporate innovation persistence and identifies key transmission mechanisms and moderating factors. Our empirical analysis reveals that extreme precipitation events significantly inhibit both innovation input and output persistence, with the impact being more pronounced on innovation outputs. This asymmetric effect suggests that environmental shocks not only disrupt resource allocation but also impair the efficiency of innovation processes themselves, particularly affecting firms’ ability to transform research inputs into valuable innovation outcomes.

The transmission mechanism analysis uncovers two primary channels through which extreme precipitation affects innovation persistence. The first channel operates through financial constraints, as extreme precipitation exacerbates firms’ financing difficulties by damaging physical assets and disrupting normal operations. This leads to increased financial pressure and reduced innovation investments, as firms must divert resources to post-disaster recovery and maintain higher liquidity reserves. The second channel works through organizational resilience, where extreme precipitation weakens firms’ ability to maintain operational stability and strategic focus. This manifests in reduced sales growth momentum and increased operational volatility, which subsequently compromises firms’ capacity to sustain consistent innovation efforts.

Our heterogeneity analysis identifies critical moderating factors that influence firms’ ability to maintain innovation persistence under extreme weather conditions. Executive characteristics play a crucial role, with R&D backgrounds and international experience significantly mitigating the negative impact of extreme precipitation. This suggests that executives with technical expertise and global exposure are better equipped to evaluate innovation projects’ strategic value and maintain research momentum during environmental disruptions. Geographic dispersion emerges as another important moderating factor, with more geographically dispersed firms demonstrating stronger resilience to extreme precipitation events, highlighting the value of spatial diversification in risk management. Ownership structure also proves to be a significant determinant of resilience to climate shocks, with non-state-owned enterprises experiencing substantially stronger negative effects from extreme precipitation compared to state-owned enterprises. This finding indicates that the superior access to government resources, implicit guarantees, and preferential policies enjoyed by state-owned enterprises provides an important buffer against environmental disruptions, allowing them to maintain more stable innovation investments even under extreme weather conditions.

Managerial implications

The findings of this study offer valuable practical implications for corporate management in an era of increasing climate uncertainty. First, companies should develop dedicated innovation continuity plans that identify critical innovation processes, potential vulnerabilities to extreme weather, and specific mitigation measures. These plans should include protocols for temporarily relocating sensitive R&D activities, protecting specialized equipment, and maintaining essential experimental conditions during extreme precipitation events.

Second, firms should consider establishing weather-derivative financial instruments to hedge against innovation disruptions from extreme precipitation. Such instruments could provide automatic liquidity injections when precipitation exceeds predetermined thresholds, ensuring continued innovation funding during challenging periods. Financial departments should integrate climate risk assessments into capital allocation decisions, potentially setting aside dedicated innovation resilience funds specifically designed to maintain R&D momentum during environmental disruptions.

Third, organizations should implement robust knowledge management systems that digitize and securely store research findings, ensuring that critical innovation insights are not vulnerable to physical damage from flooding or other extreme weather events. Cloud-based collaborative platforms can enable research continuation even when physical facilities are compromised.

Fourth, companies should consider forming innovation resilience alliances with firms in different geographic regions, establishing reciprocal agreements to temporarily share R&D facilities during extreme weather events. Such collaborative approaches could significantly reduce innovation disruptions while distributing the costs of maintaining redundant research capabilities.

Policy implications

Our findings offer several important policy implications for climate adaptation and innovation governance. First, policymakers should consider integrating precipitation risk assessments into green finance frameworks and corporate environmental, social, and governance (ESG) rating systems. By formally recognizing the impact of extreme precipitation on innovation activities, these frameworks can incentivize more climate-resilient corporate strategies and direct capital toward firms with stronger adaptive capacities.

Second, targeted innovation support mechanisms are needed for firms in regions prone to extreme precipitation events. These could include specialized R&D tax incentives, innovation infrastructure investments with enhanced weather resilience, and emergency funding mechanisms that can be rapidly deployed to maintain innovation momentum during and after extreme weather events. Such programs would help prevent innovation disruptions that could otherwise have long-lasting negative effects on regional economic development and technological advancement.

Third, government agencies should facilitate knowledge sharing and best practices for innovation resilience under climate shocks. This could involve creating public-private platforms for exchanging climate adaptation strategies specific to innovation activities, supporting the development of early warning systems tailored to innovation-intensive industries, and promoting climate-resilient innovation cluster development.

Fourth, climate adaptation policies should recognize sectoral differences in vulnerability to extreme precipitation. Our findings suggest that innovation-intensive industries with specialized equipment and precise environmental requirements (such as semiconductor manufacturing and biotechnology) may require sector-specific support measures to maintain innovation persistence under increasing climate uncertainty.

Finally, educational and workforce development policies should emphasize building human capital with both technical expertise and climate adaptation knowledge. Our findings on the moderating role of executive characteristics suggest that developing a pipeline of leaders with combined R&D and international experience could strengthen the innovation resilience of the broader economy.

Limitations and future research directions

While this study provides valuable insights into the relationship between extreme precipitation and innovation persistence, several limitations should be acknowledged. First, our instrumental variable approach using terrain slope, while theoretically sound, may not fully address all potential endogeneity concerns. Although slope primarily affects innovation through its impact on precipitation patterns, we cannot entirely rule out other channels such as historical economic development patterns tied to topography. Future research could explore alternative identification strategies, such as discontinuity designs around geographic or policy boundaries. Second, our measurement of organizational resilience, while multidimensional, may not capture all aspects of a firm’s capacity to adapt to environmental changes. The construct is inherently complex, encompassing organizational learning, leadership dynamics, and cultural elements that are challenging to quantify. Future studies could develop more comprehensive resilience measures incorporating qualitative assessments of organizational processes and capabilities. Third, while we focus on listed companies due to data availability constraints, the relationship between extreme precipitation and innovation persistence may differ for smaller, private firms that often face greater resource constraints. Extending this analysis to a broader set of companies would provide a more complete picture of how climate shocks affect the innovation ecosystem. Fourth, our study period (2012–2022) was characterized by increasing climate awareness among corporations. As climate adaptation strategies evolve, the relationships documented here may change. Longitudinal studies tracking how corporate responses to extreme precipitation evolve over longer timeframes would provide valuable insights into organizational adaptation processes. Finally, while we document the negative impact of extreme precipitation on innovation persistence, we do not fully explore potential positive adaptations, such as climate-related innovations that might emerge in response to these challenges. Future research could investigate such adaptive responses and their implications for long-term innovation trajectories.