Journal of Economic Behavior and Organization 216 (2023) 670–685

Contents lists available at ScienceDirect

Journal of Economic Behavior and Organization

journal homepage: www.elsevier.com/locate/jebo

The effect of extreme rainfall on corporate financing policies

Sicen Chen a, 1, Siyi Liu b, 1, 3, Junsheng Zhang c, d, 1, Pengdong Zhang c, d, 1, 2, *
a
Xiamen National Accounting Institute, China
b
Business School, University of International Business and Economics, China
c
School of Business, Sun Yat-sen University, China
d
Center for Accounting, Finance and Institutions, Sun Yat-sen University, China

A R T I C L E I N F O A B S T R A C T

Keywords: In recent years, a series of intense precipitation events has drawn public attention to the issue of
Extreme rainfall extreme rainfall. This study examines the impact of precipitation anomalies on corporate
Operating uncertainty financing policies and reveals that firms tend to increase their utilization of long-term debt (while
Financing policy
decreasing their reliance on short-term debt), enhance cash reserves, and reduce cash dividend
Financial resilience
payouts in response to heightened occurrences of extreme rainfall. We provide empirical evidence
indicating that such weather events can have adverse effects on firms’ operational activities. The
influence of extreme precipitation is particularly pronounced in sectors sensitive to rainfall, firms
with geographically concentrated factories, non-state-owned enterprises, and those located in
cities with limited financial institutions. Overall, this research demonstrates that firms are in­
clined to adjust their financing policies in response to extreme rainfall events, underscoring the
significance of financial resilience in managing climate risks and shedding light on firms’
financing decisions.

1. Introduction

As global temperatures continue to rise, weather-related anomalies have become increasingly significant in both practical and
academic discussions surrounding global climate change. The uneven distribution of precipitation has further exacerbated this issue,
leading to intense rainfall and floods in certain regions while others suffer from drought conditions. Recent statistical data indicates a
discernible upward trend in both the frequency and intensity of extreme rainfall events. Notably, Europe witnessed unprecedented
levels of rainfall in July 2021, resulting in severe flooding that caused extensive damage to infrastructure and disrupted communi­
cation networks. According to Liang (2022), nearly 200,000 properties experienced power outages as a consequence. Similarly, an
exceptionally heavy downpour with a maximum intensity of 201.9 mm per hour triggered devastating floods in China’s Henan
province. This event affected approximately 14.79 million individuals across 16 cities and resulted in an economic loss amounting to

* Corresponding author.
E-mail addresses: [email protected] (S. Chen), [email protected] (S. Liu), [email protected] (J. Zhang), zhangpd3@mail.
sysu.edu.cn (P. Zhang).
1
The authorship order is arranged alphabetically, with each author making an equal contribution.
2
Pengdong Zhang acknowledges the research grant from the National Natural Science Foundation of China (No. 72102242) and the Fundamental
Research Funds for the Central Universities, Sun Yat-sen University (No. 2023qntd56).
3
Siyi Liu acknowledges the research grant from the National Natural Science Foundation of China (No. 71902030), Beijing Municipal Social
Science Foundation (No. 22JCC093), and the FundamentalResearch Funds for the Central Universities in UIBE (No. CXTD14-01).

https://doi.org/10.1016/j.jebo.2023.11.005
Received 29 April 2023; Received in revised form 15 October 2023; Accepted 5 November 2023
Available online 10 November 2023
0167-2681/© 2023 Elsevier B.V. All rights reserved.
S. Chen et al. Journal of Economic Behavior and Organization 216 (2023) 670–685

12.06 billion yuan. These occurrences draw attention to the detrimental impacts associated with extreme rainfall events, which are
projected to increase according to climate models’ forecasts.
A nation or economy can suffer catastrophic losses as a result of extreme rainfall. Extensive literature demonstrates that severe
precipitation events have detrimental effects on living conditions, agricultural productivity, labor outputs, economic growth, and
social stability (e.g., Dell et al., 2014; Liang, 2022; Kotz et al., 2022). Although research specifically focusing on extreme rainfall at the
firm level is scarce, existing studies provide evidence of the adverse consequences associated with other extreme weather events such
as sea level rise and extremely high temperatures. These negative impacts include physical damage to fixed assets, reduced employee
work efficiency and safety conditions, decreased customer demand, and eventual uncertainty and volatility in business operations and
financial performance (e.g., Hong et al., 2016; Addoum et al., 2019; Park et al., 2021; Rao et al., 2021). Given the unpredictable nature
of losses caused by extreme weather conditions, firms are likely to adopt measures aimed at mitigating their adverse effects.
While insurance is commonly regarded as the primary means of protection, research has unveiled limitations to this approach,
particularly in terms of indirect losses resulting from weather events such as income loss or supply chain damage (Huang et al., 2017).
Consequently, companies cannot solely rely on insurance to mitigate the financial impacts of extreme weather conditions. Organi­
zational resilience has emerged as a crucial concept in the literature addressing the challenge of climate change. This refers to a
company’s ability to systematically absorb and recover from the adverse effects of external environmental factors, including extreme
weather conditions (Berkes et al., 2003; Tschakert and Dietrich, 2010). Previous studies have primarily focused on enhancing oper­
ational resilience towards climate change through activities relocation, infrastructure and production methods upgrading, and in­
surance coverage augmentation (Berkhout et al., 2006; Hoffmann et al., 2009). Additionally, scholars have examined the significance
of organizational slack resources encompassing backup facilities and financial reserves (Linnenluecke et al., 2008; Vogus and Sutcliffe,
2007; Woods, 2006). However further investigation is required to fully comprehend the role of financial slack in managing
climate-related risks (Linnenluecke and Griffiths, 2010).
This paper examines the impact of extreme rainfall on the financial resilience of corporations. Specifically, we analyze how extreme
rainfall affects firms’ financing policies in China. Given its status as a developing country, China faces limited access to adequate
financial instruments for hedging against extreme weather events. Consequently, firms are compelled to establish financial buffers to
mitigate potential losses caused by severe weather conditions. We propose that during periods of intense rainfall, firms are inclined
towards enhancing their financial resilience. This conjecture implies that environmental risks prompt firms to adopt conservative
financial strategies aimed at alleviating liquidity pressure and preventing financial distress.
To determine the extent of severe precipitation conditions experienced by firms, we match the rainfall data from the meteorological
station closest to a firm’s headquarters. We sort through each station’s historical diurnal observations of precipitation and select those
that exceed the 95th percentile of the series as our threshold for extreme rainfall. Subsequently, we define extreme rainfall days as
those with precipitation exceeding this threshold and measure extreme rainfall frequency as the number of such days in a calendar
year.
We assess firms’ financing strategies using four variables: (1) SHORT-TERM DEBT, representing the proportion of short-term debts
relative to lagged total assets; (2) LONG-TERM DEBT, indicating the percentage of long-term debts over lagged total assets; (3) CASH
HOLDINGS, denoting the ratio of cash, cash equivalents, and short-term investments scaled by lagged total assets; (4) CASH DIVI­
DENDS, reflecting the proportion of cash dividends scaled by lagged total assets. We hypothesize that firms exposed to more extreme
rainfall are more inclined towards utilizing long-term debts or short-term debts accordingly and maintaining higher levels of cash
holdings while distributing fewer cash dividends.
Our baseline findings support the hypothesis that firms are inclined to adopt conservative financial strategies in response to the
heightened risk of extreme rainfall. Specifically, a one-week increase in extreme rainfall days leads to (1) a 0.32 percentage point (pp)
reduction in short-term debts relative to total assets, representing a 3.4% decrease compared to the standard deviation within our
sample; (2) a 0.43 pp increase in long-term debts relative to total assets, representing a 6.83% increase compared to the standard
deviation within our sample; (3) a 0.61 pp increase in cash holdings relative to total assets, representing a 4.96% increase compared to
the standard deviation within our sample; and (4) a 0.49 pp decrease in cash dividends distribution relative to total assets, representing
a 4.05% decrease compared to the standard deviation within our sample.
To validate our inference that extreme rainfall impacts firms’ financing policies through its influence on their operations, we
construct direct measures of firms’ financial performance and operating uncertainty, subsequently examining the impact of extreme
rainfall on these variables. The findings demonstrate a reduction in firms’ returns on assets and operational cash flows due to extreme
rainfall. Furthermore, there is an observed increase in firms’ operating uncertainties, as measured by the five-year volatility of returns
on assets and operation cash flows, corresponding to higher levels of precipitation. These results provide compelling empirical evi­
dence supporting our initial inference.
We conduct multiple sets of cross-sectional regressions to enhance our inference and gain a deeper understanding of the rela­
tionship between extreme rainfall and firms’ financing policies. Firstly, we leverage the cross-sectional variation in different sectors’
susceptibility to rainfall impacts. Our rationale is that industries exhibiting higher sensitivity to excessive rain would experience a
more pronounced influence on their financing policies due to extreme rainfall events. Following Rao et al.’s (Rao et al., 2021)
approach, we define rainfall-sensitive sectors as agricultural machinery, agriculture & processed food, air transport services, tourism,
hotels & restaurants, auto sector, construction & allied activities, courier services, transport services, electricity generation & trans­
mission, fertilizers & pesticides industry and mining & quarrying sector. Our findings indicate that the association between extreme
rainfall and firms’ financing policies is stronger within these rainfall-sensitive sectors. This outcome aligns with our initial intuition.
Secondly, we hypothesize that firms with a geographically dispersed network of factories are likely to exhibit reduced vulnerability
to weather anomalies. To assess the spatial distribution of factories in terms of precipitation, we compute the average annual

671
S. Chen et al. Journal of Economic Behavior and Organization 216 (2023) 670–685

precipitation for each city. Subsequently, utilizing data from CSMAR, we determine the average precipitation coverage for each firm.
The results obtained from cross-sectional regression analysis provide support for our prediction by demonstrating that extensively
geographically distributed firms experience weakened impacts on their financing policies due to extreme rainfall events.
Additionally, we conjecture that the reliance on long-term debt and cash reserves would be diminished when a firm has access to
sufficient credit during periods of extreme rainfall. To examine this proposition, we introduce two variables: (1) STATE-OWNED
PROPERTY, denoting state-owned enterprises as one and others as zero; (2) # FINANCIAL INSTITUTIONS, representing the natural
logarithm of the number of financial institutions in a given city. Our findings indicate that the impact of extreme rainfall on firms’
financing strategies is mitigated in state-owned enterprises and cities with a higher concentration of financial institutions. These results
substantiate our hypothesis.
We also conduct several robustness tests to corroborate our inferences. Firstly, we incorporate firm fixed effects as control variables
to mitigate the influence of unobservable factors. Subsequently, we employ instrumental variable of topographic slope to address
endogeneity concerns. Moreover, we utilize the Propensity-score matching method to minimize potential biases arising from corporate
characteristics on financial policies. Lastly, to precisely match rainfall data with firms’ locations, we exclude specific samples from
Beijing, Shanghai, Guangzhou, and Shenzhen due to the likelihood of these firms operating separately from their headquarters. Our
findings remain consistent even after implementing these tests.
Our analysis contributes to the existing literature in several aspects. Firstly, our paper adds to a novel and emerging line of research
on climate risk and extreme weather events. Previous studies have primarily focused on the detrimental impact of sea level rise and
extreme temperatures on the economy and firms (Dell et al., 2014; Hong et al., 2016; Kotz et al., 2022). Some scholars have also
emphasized the influence of drought (Nguyen et al., 2022), high temperatures (Lin et al., 2019), and overall climate risks (Huang et al.,
2017) on firms’ operational and financial behaviors. In contrast, our paper goes beyond this existing research by investigating the
effect of extreme rainfall, which is one of the principal factors contributing to climate change conditions. To the best of our knowledge,
there is limited research in this domain, except for Marshall et al.’s study (Rao, 2021) examining changes in firm valuation in response
to extreme rainfall, as well as Rao et al.’s investigation (2021) into whether firms adopt distinct investment strategies under various
extreme rainfall conditions. Unlike their papers, we analyze how firms adjust their financing policies to mitigate the risk associated
with extreme rainfall.
Secondly, our paper contributes to the literature on corporate financing policies by examining the impact of extreme rainfall on
firms’ capital structure and cash holdings. While prior research has focused on factors such as management preference, control power
structure, governance mechanism, and market product competition in shaping firms’ financing decisions (e.g., Titman, 1984; Fischer
et al., 1989; Ben-Nasr et al., 2015; Cronqvist et al., 2012; Axelson et al., 2013; Lotfaliei, 2018; Zhang et al., 2022; Lin et al., 2022), there
has been less exploration of how external shocks like climate change and environmental risks affect these decisions. Although some
related studies have investigated the adverse impact of drought on corporate leverage (Nguyen et al., 2022) or comprehensive climate
risks on firms’ financing policy-making (Huang et al., 2017), our research adds to this literature by analyzing the effect of extreme
rainfall specifically. As such, it provides valuable insights for policymakers and practitioners seeking to understand corporate financial
decision-making.
The remaining sections of the paper are structured as follows. Section 2 provides a comprehensive literature review and presents
our formulated hypotheses. In Section 3, we elaborate on our sample selection process, extreme rainfall measures, methodology, and
descriptive statistics. Subsequently, Section 4 outlines our key findings along with additional analyses and robustness tests. Finally, the
concluding section encompasses our conclusions and discussions.

2. Literature review and hypotheses development

2.1. Literature on economic effects of extreme weather events

A rapidly growing body of literature examines the economic impacts of extreme weather events. Sea level rise (SLR) and extreme
temperatures have emerged as two prominent focal points in this field. SLR represents one of the most significant consequences of
climate change, posing a profound environmental risk that can greatly influence business dynamics and contribute to economic
fluctuations. Numerous studies have demonstrated that SLR risk can affect real estate and housing markets (Bernstein et al., 2019;
Filippova et al., 2019; Fuerst and Warren-Myers, 2019; Murfin and Spiegel, 2020; Fu and Jan, 2021), mortgage lending practices (Keys
and Mulder, 2020), as well as municipal bond yields (Goldsmith-Pinkham et al., 2021). While there has been extensive research on the
effects of SLR on industries and regions, limited attention has been given to corporate behaviors in response to this risk. Jiang et al.
(2019) demonstrate that the threat of SLR contributes to an increase in long-term loan spreads due to banks’ pricing in SLR risk.
Consequently, companies adjust their financial decisions by relying more on short-term loans while reducing their dependence on
long-term loans. Furthermore, Bai et al. (2020) examine how firms respond to SLR risks within the context of acquisitions. They find
that firms facing heightened exposure to SLR are more inclined to diversify this risk by acquiring entities located in areas with lower
susceptibility.
Prior research on extreme temperatures has predominantly focused on two distinct strands of inquiry. The first strand examines
whether the risk of extreme temperature is incorporated into asset prices and investigates investor perceptions of such risk. For
instance, scholars have demonstrated that environmental risks, which introduce uncertainty in business activities and cash flow, have
a negative impact on asset pricing in the market. This is evident through discounts applied to properties exposed to environmental risk
and underperformance in commercial real estate (e.g., Kumar et al., 2019; Cvijanovic and van De Minne, 2021), as well as reduced
equity valuations for corporations. Moreover, investors who recognize temperature risks exhibit a preference for avoiding such risks

672
S. Chen et al. Journal of Economic Behavior and Organization 216 (2023) 670–685

and demand compensation for environmental hazards when making investment decisions (e.g., Choi et al., 2020; Cvijanovic and van
De Minne, 2021).
Another line of research examines the impact of extreme temperature risk on corporate behavior and performance. These in­
fluences encompass various aspects, including production efficiency, safety incidents, supplier selection strategy, operational flexi­
bility, investment inclination, financing decisions, and financial performance (e.g., Addoum et al., 2019; Pankratz and Schiller, 2021).
For example, Pankratz et al. (2019) investigate how extreme temperatures affect firms’ financial performance and discover that
exposure to high temperatures can significantly diminish corporate revenues and operating income. The authors provide several
explanations for this decline such as increased cooling costs due to heat exposure, additional labor inputs and wages required to
maintain performance levels, as well as changing customer preferences. Pankratz and Schiller (2021) demonstrate that companies
adopt operational risk management strategies in their global supply chains in response to potential disruptions caused by heat.
Specifically, elevated temperatures at supplier locations can negatively impact suppliers’ and customers’ performances. Consequently,
customers tend to switch to suppliers operating in areas with lower climate-related risks to minimize such risks. Furthermore, Park
et al.(2021) emphasize that high temperatures can result in workplace accidents due to their detrimental effects on cognitive ability,
focus, and safety maintenance expenses.
Although extreme rainfall is a significant contributing factor to climate change conditions, there has been limited research con­
ducted on this topic. Among the few investigations in this area are those by Marshall et al. (2021) and Rao et al. (2021). According to
Marshall et al. (2021), rain-sensitive firms face uncertainty and potential losses due to extreme rainfall events. The authors argue that
the market may react excessively to such events, resulting in mispricing, increased trading liquidity, and heightened volatility for
related firms. In their study, Rao et al. (2021) explore how firms’ investment decisions are closely linked with opportunities arising
from extreme rainfall occurrences, indicating that companies will adjust their investments to mitigate environmental risks.

2.2. Hypotheses development

2.2.1. Effect of extreme rainfall on operating uncertainty

The impact of extreme weather on firms is multifaceted, often resulting in physical asset losses, reduced labor efficiency, opera­
tional disruptions, and decreased economic output. Previous studies have documented the adverse effects of extreme weather con­
ditions on operating costs, earnings, and productivity (Huang et al., 2017), as well as the resultant corporate uncertainty (Huang et al.,
2017; Addoum et al., 2019). In particular, Rao et al. (2021) provide direct evidence of how extreme rainfall can exacerbate uncertainty
and volatility in corporate cash flows, operating profits, and liquidity risks. These studies suggest that extreme weather conditions
increase uncertainty and volatility levels while negatively impacting financial performance across corporations.
Firms recognize that relying solely on insurance coverage is insufficient to fully shield them from the adverse consequences of
uncertainty and distress caused by extreme weather conditions (Huang et al., 2017; Jin et al., 2022). Consequently, the literature
underscores the significance of cultivating corporate resilience to climate change, which entails systematically absorbing and
recovering from the detrimental effects of external environmental disruptions (Berkes et al., 2003; Linnenluecke and Griffiths, 2010;
Tschakert and Dietrich, 2010). One effective approach to accomplish this is by augmenting organizational slack resources (e.g., Vogus
and Sutcliffe, 2007; Linnenluecke et al., 2008). Firms proactively respond to risks associated with extreme weather events through
adjustments in their investment decisions and financing choices (Huang et al., 2017; Lin et al., 2019; Nguyen et al., 2022; Marobhe and
Kansheba, 2022). As a result, building resilience and adopting an active approach can assist firms in managing the adverse impacts of
extreme weather events on their financial performance.

2.2.2. Extreme rainfall, uncertainty, and debt structure

In addition to increasing debt financing, firms will adjust their debt structure to address potential uncertainties and liquidity
pressures (Kraus and Litzenberger, 1973; Huang et al., 2017). While short-term debts are generally less costly than long-term debts, the
obligation to repay them carries greater significance. Failure to meet short-term debt obligations can expose firms to financial distress
risks and even bankruptcy. Given these concerns, companies facing liquidation risks tend to rely more heavily on long-term rather than
short-term debts (Bajaj et al., 2021). Moreover, short-term debts require frequent renegotiation of contract terms, making them more
susceptible to changes and amplifying firms’ risk of liquidation and uncertainty in financing. Consequently, a greater dependence on
long-term debts may be necessary for firms seeking to mitigate such risks and uncertainties, particularly in cases where failure to
secure sufficient funding could lead to their collapse.
As previously mentioned, extreme rainfall events are often unpredictable and can result in significant and widespread damage. This
necessitates that firms possess robust resilience to withstand such shocks (Tschakert and Dietrich, 2010). Consequently, when busi­
nesses face heightened illiquidity risk due to extreme weather conditions, they tend to adopt more cautious and conservative financing
strategies. As a result, their debt structure is predominantly characterized by long-term commitments. Based on these premises, we
propose the following hypotheses:
H1: Extreme rainfall is negatively associated with firms’ short-term debts.
H2: Extreme rainfall is positively associated with firms’ long-term debts.

2.2.3. Extreme rainfall, uncertainty, and cash holdings

Multiple studies have consistently demonstrated a strong correlation between cash holdings and dividends with agency costs,
investment opportunities, and firm governance structures. It is common for companies to maintain cash reserves to address agency
problems, seize short-term investment opportunities, and effectively manage short-term liquidity crises (Jensen, 1986; Pinkowitz

673
S. Chen et al. Journal of Economic Behavior and Organization 216 (2023) 670–685

et al., 2006; Mclean, 2011; Gao et al., 2013). Particularly during times of crisis, cash holdings play a crucial role as a safety net in
safeguarding the financial stability of the company (Opler et al., 1999; Bates et al., 2009; Duchin, 2010; Mclean, 2011).
As noted, firms encounter a range of challenges when faced with sudden extreme weather conditions, including losses in fixed
assets, employee injuries, decreased work efficiency, and diminished financial performance (Kotz et al., 2022; Liang, 2022).
Addressing these issues necessitates prompt equipment repair or replacement, compensation for employees, and restoration of pro­
ductivity. Failure to recover promptly can result in reduced operational efficiency and financial performance deterioration that may
ultimately lead to bankruptcy and liquidation. This highlights the significance of having slack resources (Woods, 2006; Linnenluecke
and Griffiths, 2010), with cash being a critical resource that enhances resilience in the face of external shocks. By directly filling gaps in
operations, investments, and financing activities, cash can help mitigate the likelihood of financial distress and liquidity crises. Taking
these factors into account leads us to anticipate an increase in firms’ cash holdings as a means to manage uncertainty and risk arising
from extreme rainfall.
Studies have demonstrated a negative correlation between the disbursement of cash dividends and corporate cash reserves,
exemplified by Easterbrook. (1984). Consequently, in periods of heavy rainfall, firms may curtail their cash dividend distributions and
bolster their liquidity buffers as a precautionary measure. Moreover, when accessing external financing becomes more challenging
(Nguyen et al., 2022), companies tend to rely more heavily on internal funding, thereby contributing to a reduction in cash dividend
payments. In summary, we contend that firms are inclined to decrease their cash dividends as a preemptive strategy to mitigate the
adverse impacts of extreme rainfall. Building upon these insights, we put forth the following hypotheses:
H3: Extreme rainfall is positively associated with cash holdings.
H4: Extreme rainfall is negatively associated with cash dividends.

3. Data and descriptive statistics

3.1. Sample selection

To investigate the impact of extreme rainfall on corporate financing policies, we initially obtained data on extreme rainfall from the
National Oceanic and Atmospheric Administration (NOAA) dataset. To create the variable of interest in this study, we extracted in­
formation on firms’ headquarters locations from the China Stock Market & Accounting Research (CSMAR) database and manually
filled in any missing values. We then converted these locations into longitude and latitude format using DBSCAN (Density-Based
Spatial Clustering of Applications with Noise).
We also retrieve financial information for firms from the CSMAR database and cross-verify it with the Wind database. The sample
period commences in 2007, coinciding with the availability of rainfall data, and concludes in 2018 prior to the advent of the COVID-19
pandemic. Finally, after excluding companies with special treatment (ST) status, companies in the financial industries, and companies
lacking necessary data for constructing control variables, we obtain a final sample consisting of 19,494 company-quarter observations.

3.2. Extreme rainfall measurement

To align the precipitation data with firm data, we initially identified the meteorological station in closest proximity to a firm’s
headquarters and established a connection between the rainfall data from this station and the listed firm. To define extreme rainfall, we
first arrange the historical diurnal observations of precipitation for each station across all years within our sample period (i.e.,
2007–2018), and consider values exceeding the 95th percentile of this series as the threshold for extreme rainfall. Subsequently, we
calculate the number of days with precipitation surpassing this threshold within a calendar year. To construct our variable of interest,
we divide the count of extreme rainfall days by 365, resulting in EXTREME RAINFALL for each observation.

3.3. Empirical design

We estimate the following regression to test whether extreme rainfall affects firms’ financing policies:

FINANCINGPOLICIESi,t = β1 EXTREMERAINFALLi,t + aind + at + r′X + εi,t (1)

where i, t, and ind index for companies, years, and industries, respectively, and FINANCING POLICIESi,q is one of the four variables: (1)
SHORT-TERM DEBT, which is the percentage of short-term debts over lagged total assets; (2) LONG-TERM DEBT, which is the per­
centage of long-term debts over lagged total assets; (3) CASH HOLDINGS, which is the percentage of cash, cash equivalent, and short-
term investment, scaled by lagged total assets; (4) CASH DIVIDENDS, which is the percentage of cash dividends scaled by lagged total
assets. EXTREME RAINFALLi,q is the frequency of extreme rainfall days in a calendar year (See Section 3.2 for a detailed process of
variable construction.)
X is a vector for controls. We control for the following firm characteristics in our analyses: the natural logarithm of total assets
(SIZE), the natural logarithm of firm age (LNAGE), intangible assets divided by lagged assets (INTANGIBLE), net property, plant, and
equipment divided by lagged assets (PPE), total liability divided by total assets (LEV), and the growth rate of sales revenue (SALES
GROWTH). To control the time-variant city characteristics, we include control variables GDP and GDP growth of the city where a firm is
located. See Appendix A for the detailed definitions of all variables. αind and αt are the industry and year indicator variables,
respectively. The industry is classified by its two-digit SIC code. We cluster standard errors by the company and winsorize all

674
S. Chen et al. Journal of Economic Behavior and Organization 216 (2023) 670–685

continuous variables at the 1st and 99th percentiles.

3.4. Descriptive statistics

Table 1 presents the descriptive statistics for the main variables used in our analyses. The average (median) for short-term debts
accounts for 9.7% (7.3%) of total assets, and the average (median) for long-term debts accounts for 4.3% (1.1%) of total assets. In
addition, the average company holds cash and cash equivalents accounting for 17.1% of total assets, and distributes cash dividends
accounting for 13.5% of total assets. Concerning rainfall conditions, the frequency of extreme rainfall days is 0.044 on average, which
means our sample is subject to an average of 16 days of extreme rainfall per year (=0.044×365). The median frequency of extreme
rainfall is 0.047, that is, the number of days affected by extreme rainfall is 17 (=0.047×365). In addition, the standard deviation of
EXTREME RAINFALL is 0.019, indicating that firms suffer different precipitation conditions.
Considering financial or market performance, Table 1 shows that the average company has a 29.6% sales growth and 43.3%
financial leverage. PPE and intangible assets account for 21.8% and 4.2% of lagged total assets, respectively. In addition, the cities
where the listed firms in our sample are located have an average GDP growth of 11.9%. These descriptive statistics are similar to the
results of earlier studies that examine Chinese company data (e.g., Pan et al., 2023; Chan et al., 2022; Liu et al., 2017).

4. Empirical results

4.1. Basic results

We first examine whether extreme rainfall affects firms’ financing policies. Table 2 presents the results of Eq. (1). Columns 1 to 4
adopt different dependent variables, i.e., SHORT-TERM DEBT, LONG-TERM DEBT, CASH HOLDINGS, and CASH DIVIDENDS,
respectively. Column 1 shows that the coefficient for EXTREME RAINFALL is negative and significant at the 1% level (coef. = − 0.165; t-
stat. = − 3.02) when the dependent variable is SHORT-TERM DEBT. The result suggests that a one-week increase in extreme rainfall
days is correlated with a 0.32% decrease (=7/365×− 0.165) in short-term debts scaled by total assets, representing a 3.4% decrease
relative to the in-sample standard deviation of SHORT-TERM DEBT. It is consistent with our hypothesis 1.
Column 2 shows that the coefficient for EXTREME RAINFALL is positive and significant at the 1% level (coef. = 0.223; t-stat. =
8.85) when the dependent variable is LONG-TERM DEBT. The result suggests that a one-week increase in extreme rainfall days is
correlated with a 0.43% increase (=7/365×0.223) in long-term debts scaled by total assets, representing a 6.83% increase relative to
the in-sample standard deviation of LONG-TERM DEBT. It is consistent with our hypothesis 2.
Column 3 shows that the coefficient for EXTREME RAINFALL is positive and significant at the 1% level (coef. = 0.318; t-stat. =
6.39) when the dependent variable is CASH HOLDINGS. The result suggests that a one-week increase in extreme rainfall days is
correlated with a 0.61% increase (=7/365×0.318) in cash holdings scaled by total assets, representing a 4.96% increase relative to the
in-sample standard deviation of CASH HOLDINGS. It is consistent with our hypothesis 3.
Column 4 shows that the coefficient for EXTREME RAINFALL is negative and significant at the 1% level (coef. = − 0.255; t-stat. =
− 3.54) when the dependent variable is CASH DIVIDENDS. The result suggests that a one-week increase in extreme rainfall days is
correlated with a 0.49% decrease (=7/365×− 0.255) in cash dividends scaled by total assets, representing a 4.05% increase relative to
the in-sample standard deviation of CASH DIVIDENDS. It is consistent with our hypothesis 4.
Overall, the results in Table 2 document that firms experiencing extreme rainfall are more likely to use long-term debts and less
likely to take short-term debts. In addition, they will hold more cash and distribute less cash dividends. These results support all our
hypotheses.

Table 1
Descriptive statistics.
variable Mean SD P25 P50 P75 N

SHORT-TERM DEBT 0.097 0.094 0.012 0.073 0.157 19,494

LONG-TERM DEBT 0.043 0.063 0.000 0.011 0.060 19,494
CASH HOLDINGS 0.171 0.123 0.079 0.135 0.230 19,494
CASH DIVIDENDS 0.135 0.121 0.050 0.100 0.180 19,494
ROA 0.042 0.040 0.016 0.038 0.066 19,494
ROA VOLATILITY 0.018 0.011 0.008 0.015 0.025 19,494
OCF 0.043 0.063 0.004 0.043 0.085 19,494
OCF VOLATILITY 0.034 0.021 0.018 0.030 0.046 19,494
EXTREME RAINFALL 0.044 0.019 0.036 0.047 0.055 19,494
SIZE 20.118 1.526 19.097 20.018 21.085 19,494
LNAGE 2.002 0.873 1.386 2.197 2.708 19,494
INTANGIBLE 0.042 0.036 0.015 0.032 0.057 19,494
LEVERAGE 0.433 0.203 0.266 0.427 0.592 19,494
PPE 0.218 0.158 0.090 0.186 0.318 19,494
SALES GROWTH 0.296 0.545 − 0.032 0.134 0.429 19,494
LNGDP 17.828 1.062 17.038 17.933 18.735 19,494
GDP GROWTH 0.119 0.050 0.083 0.112 0.154 19,494

675
S. Chen et al. Journal of Economic Behavior and Organization 216 (2023) 670–685

Table 2
Extreme rainfall and financing policies.
Dependent Variables: SHORT-TERM DEBT LONG-TERM DEBT CASH HOLDINGS CASH DIVIDENDS
(1) (2) (3) (4)
Coefficient Coefficient Coefficient Coefficient
t-Statistic t-Statistic t-Statistic t-Statistic

EXTREME RAINFALL − 0.165*** 0.223*** 0.318*** − 0.255***

(− 3.02) (8.85) (6.39) (− 3.54)
SIZE 0.008*** 0.012*** 0.004*** 0.026***
(6.65) (15.38) (2.86) (12.41)
LNAGE 0.013*** 0.003*** − 0.023*** − 0.033***
(8.58) (3.36) (− 13.80) (− 15.56)
INTANGIBLE 0.110*** 0.000 − 0.332*** − 0.164***
(2.87) (0.01) (− 9.21) (− 3.62)
PPE 0.054*** 0.029*** − 0.253*** − 0.173***
(4.17) (3.76) (− 22.08) (− 10.73)
SALES GROWTH − 0.002 0.006*** − 0.001 − 0.007***
(− 1.28) (4.45) (− 0.30) (− 3.31)
LNGDP − 0.004*** − 0.001 0.003* − 0.001
(− 2.86) (− 0.81) (1.87) (− 0.38)
GDP GROWTH 0.004 0.053*** 0.092*** − 0.025
(0.15) (3.35) (3.29) (− 0.75)
LEVERAGE − 0.225*** − 0.080***
(− 26.68) (− 7.17)
Industry Fixed Effects Included Included Included Included
Year Fixed Effects Included Included Included Included
Adj. R-squared 0.197 0.365 0.427 0.150
Observations 19,494 19,494 19,494 19,494

Notes: This table presents the results from regressions of financing policy proxies on extreme rainfall and control variables. See Appendix A for
variable definitions. The t-statistics are clustered at the company level. ***, **, and * denote statistical significance at the two-tailed 1%, 5%, and 10%
levels, respectively.

Table 3
Extreme rainfall and firms’ operation.
Dependent Variables: ROA OCF ROA VOLATILITY OCF VOLATILITY
(1) (2) (3) (4)
Coefficient Coefficient Coefficient Coefficient
t-Statistic t-Statistic t-Statistic t-Statistic

EXTREME RAINFALL − 0.063*** − 0.086*** 0.007* 0.020**

(− 3.27) (− 3.08) (1.82) (2.34)
ROA 0.164***
(41.95)
OCF 0.170***
(45.06)
SIZE 0.007*** 0.006*** − 0.001*** − 0.002***
(14.48) (9.11) (− 12.25) (− 11.06)
LNAGE − 0.006*** 0.002*** 0.001*** − 0.000
(− 10.67) (2.85) (7.60) (− 0.01)
INTANGIBLE − 0.060*** 0.061*** 0.005 − 0.025***
(− 4.50) (3.22) (1.58) (− 4.33)
LEVERAGE − 0.084*** − 0.061*** 0.002*** 0.014***
(− 28.22) (− 15.34) (3.00) (10.71)
PPE − 0.065*** 0.068*** 0.008*** − 0.006***
(− 15.30) (11.09) (8.23) (− 2.88)
SALES GROWTH 0.001* − 0.007*** − 0.000 0.003***
(1.82) (− 6.49) (− 0.31) (8.36)
LNGDP 0.000 0.001 0.000 − 0.000
(0.29) (0.93) (0.92) (− 0.48)
GDP GROWTH 0.021** 0.018 − 0.006** − 0.001
(2.01) (1.14) (− 2.26) (− 0.20)
Industry Fixed Effects Included Included Included Included
Year Fixed Effects Included Included Included Included
Adj. R-squared 0.256 0.173 0.374 0.296
Observations 19,494 19,494 19,494 19,494

Notes: This table presents the results from regressions of firms’ financial performance and operating uncertainty on extreme rainfall and control
variables. See Appendix A for variable definitions. The t-statistics are clustered at the company level. ***, **, and * denote statistical significance at
the two-tailed 1%, 5%, and 10% levels, respectively.

676
S. Chen et al. Journal of Economic Behavior and Organization 216 (2023) 670–685

Table 4
Cross-sectional analyses from the perspective of rainfall
Panel A: Cross-sectional analyses of rainfall-sensitive sectors.
Dependent Variables: SHORT-TERM DEBT LONG-TERM DEBT CASH HOLDINGS CASH DIVIDENDS
(1) (2) (3) (4)
Coefficient Coefficient Coefficient Coefficient
t-Statistic t-Statistic t-Statistic t-Statistic

EXTREME RAINFALL − 0.115** 0.190*** 0.286*** − 0.199***

(− 1.98) (7.62) (5.23) (− 2.67)
× RAIN-SENSITIVE INDUSTRIES − 0.386** 0.247*** 0.243** − 0.430**
(− 2.46) (3.15) (2.14) (− 1.96)
SIZE 0.008*** 0.012*** 0.004*** 0.026***
(6.64) (15.39) (2.86) (12.43)
LNAGE 0.013*** 0.003*** − 0.023*** − 0.033***
(8.56) (3.40) (− 13.79) (− 15.59)
INTANGIBLE 0.111*** − 0.000 − 0.332*** − 0.163***
(2.89) (− 0.01) (− 9.22) (− 3.60)
PPE 0.054*** 0.029*** − 0.253*** − 0.173***
(4.21) (3.72) (− 22.09) (− 10.73)
SALES GROWTH − 0.002 0.006*** − 0.001 − 0.007***
(− 1.33) (4.48) (− 0.27) (− 3.34)
LNGDP − 0.004*** − 0.001 0.003* − 0.001
(− 2.83) (− 0.84) (1.85) (− 0.36)
GDP GROWTH 0.004 0.053*** 0.092*** − 0.026
(0.13) (3.36) (3.30) (− 0.76)
LEVERAGE − 0.225*** − 0.080***
(− 26.65) (− 7.19)
Industry Fixed Effects Included Included Included Included
Year Fixed Effects Included Included Included Included
Adj. R-squared 0.198 0.365 0.427 0.151
Observations 19,494 19,494 19,494 19,494

Panel B: Cross-sectional analyses of factories’ geographical distribution

Dependent Variables: SHORT-TERM DEBT LONG-TERM DEBT CASH HOLDINGS CASH DIVIDENDS
(1) (2) (3) (4)
Coefficient Coefficient Coefficient Coefficient
t-Statistic t-Statistic t-Statistic t-Statistic

EXTREME RAINFALL − 0.223*** 0.256*** 0.384*** − 0.353***

(− 3.24) (9.00) (6.45) (− 3.77)
× WIDELY DISTRIBUTED 0.187* − 0.109** − 0.211** 0.321***
(1.88) (− 2.08) (− 2.15) (2.65)
WIDELY DISTRIBUTED − 0.012** 0.009*** 0.006 − 0.024***
(− 2.24) (3.19) (1.13) (− 3.72)
SIZE 0.008*** 0.012*** 0.004** 0.025***
(6.03) (15.10) (2.45) (11.37)
LNAGE 0.013*** 0.003*** − 0.023*** − 0.033***
(8.65) (3.23) (− 13.71) (− 15.36)
INTANGIBLE 0.110*** 0.001 − 0.333*** − 0.164***
(2.87) (0.02) (− 9.24) (− 3.64)
PPE 0.054*** 0.029*** − 0.253*** − 0.173***
(4.20) (3.72) (− 22.11) (− 10.71)
SALES GROWTH − 0.002 0.006*** − 0.001 − 0.007***
(− 1.19) (4.33) (− 0.28) (− 3.13)
LNGDP − 0.004*** − 0.001 0.003* − 0.001
(− 2.80) (− 0.89) (1.85) (− 0.29)
GDP GROWTH 0.003 0.054*** 0.093*** − 0.028
(0.10) (3.41) (3.31) (− 0.82)
LEVERAGE − 0.226*** − 0.080***
(− 26.66) (− 7.24)
Industry Fixed Effects Included Included Included Included
Year Fixed Effects Included Included Included Included
Adj. R-squared 0.198 0.366 0.427 0.152
Observations 19,494 19,494 19,494 19,494

Notes: This table presents the results of cross-sectional analyses from the perspective of rainfall. The regressions in Panel A estimate the impact of
rainfall sensitivity on the relationship between extreme rainfall and financing policies, while the regressions in Panel B estimate the impact of the
geographical distribution of factories on the relationship between extreme rainfall and financing policies. See Appendix A for variable definitions. The
t-statistics are clustered at the company level. ***, **, and * denote statistical significance at the two-tailed 1%, 5%, and 10% levels, respectively.

677
S. Chen et al. Journal of Economic Behavior and Organization 216 (2023) 670–685

4.2. Effect of extreme rainfall on firms’ financial performance and operating uncertainty

We conjecture that extreme precipitation has a negative (positive) effect on firms’ financial performance (operational uncertainty),
prompting them to adjust their financing policies accordingly. To validate our conjecture, we construct direct measures of firms’
financial performance and operational uncertainty and then investigate the impact of extreme precipitation on these variables.
First, we use return on assets (ROA) and operating cash flows (OCF) to measure firms’ financial performance. We define ROA as net
income divided by total assets, and OCF as operating cash flows divided by total assets. Then, we replace the dependent variable in Eq.
(1) with these two performance variables, ROA and OCF, and rerun the regressions. Columns 1 and 2 of Table 3 present the results.
Column 1 shows that the coefficient for EXTREME RAINFALL is negative and significant at the 1% level (coef. = − 0.063; t-stat. =
− 3.27) when the dependent variable is ROA. The result suggests that a one-week increase in extreme rainfall days is correlated with a
0.12% decrease in return on assets, representing a 3% decrease relative to the in-sample standard deviation of ROA. Column 2 of
Table 3 shows that the coefficient for EXTREME RAINFALL is negative and significant at the 1% level (coef. = − 0.086; t-stat. = − 3.08)
when the dependent variable is OCF. The result suggests that a one-week increase in extreme rainfall days is correlated with a 0.16%
decrease in operating cash flows scaled by lagged total assets, representing a 2.54% decrease relative to the in-sample standard de­
viation of OCF. Both results in columns 1 and 2 are consistent with our conjecture.
Second, we create two new variables to measure firms’ operating uncertainty: (1) ROA VOLATILITY, which is the standard de­
viation of quarterly pre-tax income scaled by total assets over the preceding five fiscal years; (2) OCF VOLATILITY, which is measured
by the standard deviation of quarterly cash flows from operations scaled by total assets over the preceding five fiscal years. We replace
the dependent variables in Eq. (1) with these two performance variables, ROA VOLATILITY and OCF VOLATILITY, and rerun the
regressions. Note that we follow the existing literature and add ROA (OCF) into controls when the dependent variable is ROA
VOLATILITY (OCF VOLATILITY). Columns 3 and 4 of Table 3 present the results. Column 3 shows that the coefficient for EXTREME
RAINFALL is positive and significant at the 10% level (coef. = 0.007; t-stat. = 1.82) when the dependent variable is ROA VOLATILITY.
The result suggests that a one-week increase in extreme rainfall days is correlated with a 0.01% increase in five-year volatility of return
on assets, representing a 0.9% increase relative to the in-sample standard deviation of ROA VOLATILITY. Column 4 of Table 3 shows
that the coefficient for EXTREME RAINFALL is positive and significant at the 5% level (coef. = 0.020; t-stat. = 2.34) when the
dependent variable is OCF VOLATILITY. The result suggests that a one-week increase in extreme rainfall days is correlated with a
0.04% increase in five-year volatility of operating cash flows, representing a 1.9% increase relative to the in-sample standard deviation
of OCF VOLATILITY. Both results in columns 3 and 4 are consistent with our conjecture.
Overall, the results in Table 3 provide evidence supporting that extreme rainfall reduces firms’ financial performance and increases
firms’ operating uncertainty. This is consistent with our inference that extreme rainfall leads to firms’ change in financing policies due
to its effect on firms’ physical operations.

4.3. Cross-sectional analyses

4.3.1. The perspective of rainfall

We predict that industries that are much more sensitive to excess rain would witness a more pronounced effect of extreme rainfall
on firms’ financing policies. Following Rao et al. (2021), we define rainfall-sensitive sectors as agricultural machinery, agriculture &
processed food, air transport services, tourism, hotels & restaurants, auto sector, construction & allied activities, courier services,
transport services, electricity generation & transmission, fertilizers & pesticides, and mining & quarrying. We follow their way and
construct a dummy variable, RAIN-SENSITIVE INDUSTRIES, which equals one for these rain-sensitive industries and zero otherwise.
We then estimate a modified version of Eq. (1) that includes an interaction term between RAIN-SENSITIVE INDUSTRIES and
EXTREME RAINFALL. In Panel A of Table 4, columns 1 to 4 present the regression results of the modified equation for different
dependent variables. Column 1 shows that the coefficient for EXTREME RAINFALL×RAIN-SENSITIVE INDUSTRIES is negative and
significant at the 5% level (coef. = − 0.386; t-stat. = − 2.46), indicating that firms in rainfall-sensitive sectors are more likely to reduce
the use of short-term debts when they experience severer extreme rainfall. Column 2 shows that the coefficient for EXTREME
RAINFALL×RAIN-SENSITIVE INDUSTRIES is positive and significant at the 1% level (coef. = 0.247; t-stat. = 3.15), indicating that the
positive effect of extreme rainfall on firms’ long-term debts is more pronounced in rainfall-sensitive sectors. Column 3 shows that the
coefficient for EXTREME RAINFALL×RAIN-SENSITIVE INDUSTRIES is positive and significant at the 5% level (coef. = 0.243; t-stat. =
2.14), indicating that firms in rainfall-sensitive sectors are more likely to hold cash and cash equivalents when they experience severer
extreme rainfall. Column 4 shows that the coefficient for EXTREME RAINFALL×RAIN-SENSITIVE INDUSTRIES is negative and sig­
nificant at the 5% level (coef. = − 0.430; t-stat. = − 1.96), indicating that the negative effect of extreme rainfall on firms’ distribution of
cash dividends is more pronounced in rainfall-sensitive sectors. Taken together, the combined results are consistent with our pre­
diction that the effect of extreme rainfall on firms’ financing policies would be stronger in rainfall-sensitive sectors, which are
consistent with our prediction.
Since extreme rainfall events are localized, another inference is that firms whose factories are widely geographically distributed are
less likely to be affected by weather anomalies. We create a new variable, WIDELY DISTRIBUTED, to measure the geographical dis­
tribution of factories for a firm from the perspective of precipitation. Specifically, we calculate the average precipitation for each city
year. Then, we get factory information from the CSMAR and calculate the average precipitation coverage for different factories of a
firm.
We then estimate a modified version of Eq. (1) that includes an interaction term between the variable, WIDELY DISTRIBUTED, and
EXTREME RAINFALL. In Panel B of Table 4, columns 1 to 4 present the regression results of the modified equation for different

678
S. Chen et al. Journal of Economic Behavior and Organization 216 (2023) 670–685

Table 5
Cross-sectional analyses from the perspective of financing
Panel A: Cross-sectional analyses of state-owned property.
Dependent Variables: SHORT-TERM DEBT LONG-TERM DEBT CASH HOLDINGS CASH DIVIDENDS
(1) (2) (3) (4)
Coefficient Coefficient Coefficient Coefficient

t-Statistic t-Statistic t-Statistic t-Statistic

EXTREME RAINFALL − 0.284*** 0.255*** 0.345*** − 0.329***
(− 4.19) (9.36) (5.85) (− 3.72)
× STATE-OWNED PROPERTY 0.445*** − 0.095* − 0.214** 0.246*
(4.77) (− 1.66) (− 2.30) (1.80)
STATE-OWNED PROPERTY − 0.031*** 0.003 0.033*** − 0.012
(− 6.07) (0.88) (6.72) (− 1.57)
SIZE 0.009*** 0.012*** 0.003* 0.026***
(7.19) (15.50) (1.82) (12.47)
LNAGE 0.015*** 0.003*** − 0.027*** − 0.033***
(9.53) (3.39) (− 15.64) (− 13.99)
INTANGIBLE 0.103*** − 0.000 − 0.319*** − 0.165***
(2.71) (− 0.01) (− 8.88) (− 3.65)
PPE 0.053*** 0.030*** − 0.252*** − 0.173***
(4.15) (3.77) (− 21.95) (− 10.75)
SALES GROWTH − 0.002 0.006*** − 0.001 − 0.007***
(− 1.22) (4.49) (− 0.46) (− 3.32)
LNGDP − 0.004*** − 0.001 0.002 − 0.001
(− 2.58) (− 0.73) (1.25) (− 0.37)
GDP GROWTH 0.008 0.054*** 0.082*** − 0.025
(0.31) (3.40) (2.94) (− 0.75)
LEVERAGE − 0.227*** − 0.080***
(− 26.86) (− 7.18)

Industry Fixed Effects Included Included Included Included

Year Fixed Effects Included Included Included Included
Adj. R-squared 0.201 0.365 0.432 0.150
Observations 19,494 19,494 19,494 19,494

Panel B: Cross-sectional analyses of # financial institutions

Dependent Variables: SHORT-TERM DEBT LONG-TERM DEBT CASH HOLDINGS CASH DIVIDENDS
(1) (2) (3) (4)
Coefficient Coefficient Coefficient Coefficient
t-Statistic t-Statistic t-Statistic t-Statistic

EXTREME RAINFALL − 0.392*** 0.282*** 0.448*** − 0.543***

(− 5.72) (8.35) (5.44) (− 4.05)
× # FINANCIAL INSTITUTIONS 0.348*** − 0.085* − 0.196* 0.433***
(3.60) (− 1.83) (− 1.94) (2.83)
# FINANCIAL INSTITUTIONS − 0.006 0.006** 0.007 − 0.015
(− 0.90) (2.18) (1.08) (− 1.60)
SIZE 0.008*** 0.012*** 0.004*** 0.026***
(6.67) (15.42) (2.85) (12.46)
LNAGE 0.013*** 0.003*** − 0.023*** − 0.033***
(8.63) (3.33) (− 13.83) (− 15.50)
INTANGIBLE 0.111*** 0.001 − 0.332*** − 0.163***
(2.91) (0.03) (− 9.23) (− 3.60)
PPE 0.055*** 0.030*** − 0.253*** − 0.173***
(4.28) (3.79) (− 22.15) (− 10.69)
SALES GROWTH − 0.003 0.006*** − 0.001 − 0.008***
(− 1.41) (4.48) (− 0.25) (− 3.41)
LNGDP − 0.008*** − 0.002 0.003 − 0.002
(− 3.27) (− 1.36) (1.54) (− 0.71)
GDP GROWTH − 0.017 0.051*** 0.099*** − 0.040
(− 0.63) (3.22) (3.53) (− 1.18)
LEVERAGE − 0.225*** − 0.080***
(− 26.61) (− 7.25)
Industry Fixed Effects Included Included Included Included
Year Fixed Effects Included Included Included Included
Adj. R-squared 0.199 0.365 0.427 0.151
Observations 19,494 19,494 19,494 19,494

Notes: This table presents the results of cross-sectional analyses from the perspective of financing. The regressions in Panel A estimate the impact of
state-owned property on the relationship between extreme rainfall and financing policies, while the regressions in Panel B estimate the impact of #
financial institutions on the relationship between extreme rainfall and financing policies. See Appendix A for variable definitions. The t-statistics are
clustered at the company level. ***, **, and * denote statistical significance at the two-tailed 1%, 5%, and 10% levels, respectively.

679
S. Chen et al. Journal of Economic Behavior and Organization 216 (2023) 670–685

dependent variables. Column 1 shows that the coefficient for EXTREME RAINFALL×WIDELY DISTRIBUTED is positive and significant
at the 10% level (coef. = 0.187; t-stat. = 1.88), indicating that firms with widely geographically distributed factories are more likely to
borrow more short-term debts when they experience severer extreme rainfall. Column 2 shows that the coefficient for EXTREME
RAINFALL×WIDELY DISTRIBUTED is negative and significant at the 5% level (coef. = − 0.109; t-stat. = − 2.08), indicating that the
effect of extreme rainfall on long-term debts is weakened in firms with widely geographically distributed factories. Column 3 shows
that the coefficient for EXTREME RAINFALL×WIDELY DISTRIBUTED is negative and significant at the 5% level (coef. = − 0.211; t-stat.
= − 2.15), indicating that firms with narrowly geographically distributed factories are more likely to hold cash and cash equivalents
when they experience severer extreme rainfall. Column 4 shows that the coefficient for EXTREME RAINFALL×WIDELY DISTRIBUTED
is positive and significant at the 1% level (coef. = 0.321; t-stat. = 2.65), indicating that the effect of extreme rainfall on firms’ dis­
tribution of cash dividends is less pronounced in firms with narrowly geographically distributed factories. Taken together, the com­
bined results are consistent with our prediction that the effect of extreme rainfall on firms’ financing policies would be weakened in
widely geographically distributed firms, which is consistent with our prediction.

4.3.2. The perspective of financing

We conjecture that firms experiencing extreme rainfall will extend their debt duration and retain more cash for precautionary
purposes. However, the need for long-term debt and cash holdings would be weakened when a firm is available to get enough credit in
the case of need. We adopt two variables to describe this status: (1) STATE-OWNED PROPERTY, which equals one for state-owned
enterprises, and zero otherwise; (2) # FINANCIAL INSTITUTIONS, which is the natural logarithm of the number of financial in­
stitutions in a city.
In Panel A of Table 5, we estimate a modified version of Eq. (1) that includes an interaction term between STATE-OWNED
PROPERTY and EXTREME RAINFALL. Column 1 shows that the coefficient for EXTREME RAINFALL×STATE-OWNED PROPERTY is
positive and significant at the 1% level (coef. = 0.445; t-stat. = 4.77), indicating that state-owned enterprises are more likely to borrow
more short-term debts when they experience severer extreme rainfall. Column 2 shows that the coefficient for EXTREME RAIN­
FALL×STATE-OWNED PROPERTY is negative and significant at the 10% level (coef. = − 0.095; t-stat. = − 1.66), indicating that the
effect of extreme rainfall on long-term debts is weakened in state-owned enterprises. Column 3 shows that the coefficient for EXTREME
RAINFALL×STATE-OWNED PROPERTY is negative and significant at the 5% level (coef. = − 0.214; t-stat. = − 2.30), indicating that
state-owned enterprises are more likely to hold cash and cash equivalents when they experience severer extreme rainfall. Column 4
shows that the coefficient for EXTREME RAINFALL×STATE-OWNED PROPERTY is positive and significant at the 10% level (coef. =
0.246; t-stat. = 1.80), indicating that the effect of extreme rainfall on firms’ distribution of cash dividends is less pronounced in state-
owned enterprises. Taken together, the combined results are consistent with our prediction that the effect of extreme rainfall on firms’
financing policies would be weakened in state-owned enterprises, which are consistent with our prediction.
In Panel B of Table 5, we estimate a modified version of Eq. (1) that includes an interaction term between # FINANCIAL IN­
STITUTIONS and EXTREME RAINFALL. Column 1 shows that the coefficient for EXTREME RAINFALL×# FINANCIAL INSTITUTIONS is

Table 6
Regressions with firm-fixed effects.
Dependent Variables: SHORT-TERM DEBT LONG-TERM DEBT CASH HOLDINGS CASH DIVIDENDS
(1) (2) (3) (4)
Coefficient Coefficient Coefficient Coefficient
t-Statistic t-Statistic t-Statistic t-Statistic

EXTREME RAINFALL − 0.207*** 0.049* 0.093* − 0.294***

(− 5.29) (1.75) (1.87) (− 5.96)
SIZE 0.010*** 0.015*** 0.000 0.003*
(5.44) (12.24) (0.03) (1.92)
LNAGE 0.028*** 0.003** − 0.095*** − 0.063***
(12.43) (2.29) (− 31.57) (− 20.42)
INTANGIBLE 0.079** 0.025 − 0.408*** − 0.127***
(2.11) (0.98) (− 10.34) (− 4.18)
PPE 0.030** − 0.040*** − 0.251*** − 0.054***
(2.23) (− 4.66) (− 17.19) (− 4.87)
SALES GROWTH − 0.004*** 0.005*** − 0.003** − 0.002
(− 3.23) (4.60) (− 2.08) (− 1.19)
LNGDP 0.008 0.009** − 0.008 0.002
(1.24) (2.24) (− 1.29) (0.60)
GDP GROWTH − 0.002 0.018 − 0.002 − 0.030
(− 0.08) (1.42) (− 0.10) (− 1.40)
LEVERAGE − 0.200*** − 0.045***
(− 22.56) (− 6.26)
Company Fixed Effects Included Included Included Included
Year Fixed Effects Included Included Included Included
Adj. R-squared 0.643 0.647 0.679 0.642
Observations 19,494 19,494 19,494 19,494

Notes: This table presents the results of robustness tests by controlling firm fixed effects. See Appendix A for variable definitions. The t-statistics are
clustered at the company level. ***, **, and * denote statistical significance at the two-tailed 1%, 5%, and 10% levels, respectively.

680
S. Chen et al. Journal of Economic Behavior and Organization 216 (2023) 670–685

positive and significant at the 1% level (coef. = 0.348; t-stat. = 3.60), indicating that firms located in cities with more financial in­
stitutions are more likely to borrow more short-term debts when they experience severer extreme rainfall. Column 2 shows that the
coefficient for EXTREME RAINFALL×# FINANCIAL INSTITUTIONS is negative and significant at the 10% level (coef. = − 0.085; t-stat.
= − 1.83), indicating that the effect of extreme rainfall on long-term debts is weakened in firms located in cities with more financial
institutions. Column 3 shows that the coefficient for EXTREME RAINFALL×# FINANCIAL INSTITUTIONS is negative and significant at
the 10% level (coef. = − 0.196; t-stat. = − 1.94), indicating that firms located in cities with more financial institutions are more likely to
hold cash and cash equivalents when they experience severer extreme rainfall. Column 4 shows that the coefficient for EXTREME
RAINFALL×# FINANCIAL INSTITUTIONS is positive and significant at the 1% level (coef. = 0.433; t-stat. = 2.83), indicating that the
effect of extreme rainfall on firms’ distribution of cash dividends is less pronounced in firms located in cities with more financial
institutions. Taken together, the combined results are consistent with our prediction that the effect of extreme rainfall on firms’
financing policies would be weakened in cities with more financial institutions, which is consistent with our prediction.

4.4. Robustness tests

4.4.1. Include firm indicator variables

Existing literature documents that a firm’s financing policies hold constantly for a long period. Meanwhile, the headquarters of
most firms are unchanged in our sample period. Given so, the issue of omitted variables arises since there might be firm-specific
characteristics that influence both firms’ financing policies and their location selections. To address the issue of endogeneity, we
have incorporated firm-fixed effects in all our models as a means of mitigation. Firm-fixed effects control for time-invariant firm
characteristics that may affect both a firm’s decisions on financing and its selection of headquarters location. By incorporating firm-
fixed effects, our findings can be interpreted as being specific to within-firm analysis.
Table 6 presents the regression results of Eq. (1) when replacing industry-fixed effects with firm-fixed effects. The results in column
1 (column 2) show that the coefficient of EXTREME RAINFALL is negative (positive) when the dependent variable is SHORT-TERM
DEBT (LONG-TERM DEBT), which is consistent with hypothesis 1 (hypothesis 2). Column 3 (column 4) shows that the coefficient of
EXTREME RAINFALL is positive (negative) when the dependent variable is CASH HOLDINGS (CASH DIVIDENDS), which is consistent
with hypothesis 3 (hypothesis 4). Overall, the results in Table 6 when adding firm-fixed effects support our hypotheses.

4.4.2. Instrument variable method

To further address the endogeneity, we also examine the relationship between extreme rainfall and financial policies by the
instrumental variable approach. According to our research problem, effective instrumental variables must meet two requirements: (1)
Instrumental variables are correlated with extreme rainfall, which indicates correlation; (2) Instrumental variables have nothing to do

Table 7
Results of the instrumental variable approach.
Dependent Variables: EXTREME RAINFALL SHORT-TERM DEBT LONG-TERM DEBT CASH HOLDINGS CASH DIVIDENDS
(1) (2) (3) (4) (5)
Coefficient Coefficient Coefficient Coefficient Coefficient
t-Statistic t-Statistic t-Statistic t-Statistic t-Statistic

SLOPE − 0.000***
(− 8.94)
EXTREME RAINFALL − 0.976*** 1.463*** 2.209*** − 3.442***
(− 2.86) (2.63) (3.58) (− 4.39)
SIZE 0.000 0.013*** 0.008*** 0.003*** 0.027***
(1.63) (30.08) (11.14) (4.58) (27.98)
LNAGE − 0.001*** 0.002*** 0.014*** − 0.022*** − 0.036***
(− 3.90) (4.02) (15.70) (− 20.01) (− 26.17)
INTANGIBLE 0.007* 0.009 0.098*** − 0.347*** − 0.139***
(1.74) (0.74) (4.81) (− 15.36) (− 4.86)
PPE 0.004*** − 0.233*** − 0.067***
(4.84) (− 44.84) (− 10.17)
SALES GROWTH − 0.002* 0.026*** 0.058*** − 0.249*** − 0.180***
(− 1.94) (6.56) (8.99) (− 34.77) (− 19.90)
LNGDP 0.000 0.006*** − 0.003** − 0.001 − 0.007***
(0.31) (7.60) (− 2.05) (− 0.54) (− 3.76)
GDP GROWTH − 0.003*** − 0.004*** 0.001 0.009*** − 0.011***
(− 22.23) (− 3.84) (0.32) (4.17) (− 4.06)
LEVERAGE 0.012*** 0.067*** − 0.015 0.071*** 0.009
(3.12) (5.26) (− 0.71) (3.12) (0.32)
Industry Fixed Effects Included Included Included Included Included
Year Fixed Effects Included Included Included Included Included
Adj. R-squared 0.085 0.051 0.056 0.203 0.143
Observations 19,494 19,494 19,494 19,494 19,494

Notes: This table presents the results of the instrumental variable approach. See Appendix A for variable definitions. The t-statistics are clustered at the
company level. ***, **, and * denote statistical significance at the two-tailed 1%, 5%, and 10% levels, respectively.

681
S. Chen et al. Journal of Economic Behavior and Organization 216 (2023) 670–685

with corporate financing decisions, that is, exogenous.

Research has indicated that elevated temperatures can have an impact on precipitation patterns. White et al. (2022) also document
the effect of topographic conditions on the development of extreme rainfall. Both high temperature and topographic slope are related
to extreme rainfall. Therefore, they are valid instrumental variables for extreme rainfall, which are used as instrumental variables of
extreme rainfall in our paper. Therefore, we construct an instrument variables, SLOPE. SLOPE is computed by first dividing each city’s
administrative area into ten square km grids based on longitude and latitude data. Next, calculate the difference between the maximum
and minimum elevations for each grid. The SLOPE for each city is the average of these differences.
Table 7 presents the regression outcomes of the instrumental variable approach. In the initial stage of regression, we adopt extreme
rainfall as the dependent variable and conduct probit regression on the instrumental variables and control variables in the baseline
model. The findings demonstrate that slope coefficients are significant, indicating its effectiveness as instrumental variables. In the
second stage, we consider the results of probit regression in the first stage to investigate the influence of extreme rainfall on corporate
financial decisions. The IV regressions successfully pass both the weak instrumental variable test and the over-identification test. As
shown, our main results are unaltered.

4.4.3. Propensity-score matching

Corporate characteristics, such as firm size, operating performance, and corporate governance, can influence firms’ financial
policies. Hence, the differences in financial policies may be attributed to these characteristics rather than extreme rainfall. To address
the self-selection concern, we employ the propensity-score matching (PSM) approach. Referring to Rubin (1973) and Dehejia and
Wahba (2002), through pairing treatment and control samples, the PSM approach can mitigate some selection bias arising from
observable factors. Accordingly, employing this approach, we generate a binary variable to distinguish the samples into the treatment
group and the control group based on the seventy-fifth percentile of EXTREME RAINFALL. The treatment group comprises observations
with EXTREME RAINFALL values higher than or equal to the seventy-fifth percentile, assigned a value of 1. Conversely, observations
with EXTREME RAINFALL values lower than the seventy-fifth percentile are allocated to the potential control group, denoted by a
dummy variable set at 0.
A logit regression is then employed to estimate the likelihood of a firm receiving treatment, using all control variables from the
baseline model. Subsequently, we compute the propensity score for each treated firm based on predicted probabilities derived from the
logit model. Finally, we employ radius matching with replacement to match each treated firm with a control firm. In so doing, we use a
maximum allowable range of propensity score radius of 0.01. To evaluate the effectiveness of matching, we compare the average level
of the covariates in Eq. (1) between treatment and control samples, and no statistically significant difference was observed between the
treatment and control groups in all dimensions of the covariates following matching. This suggests that our matching procedure
enhances covariate balance. Then, we utilize this PSM sample to re-estimate Eq. (1). Table 8 shows the PSM results, which are similar
to our baseline regression results. Overall, our findings remain unaltered when we account for potential selection bias.

Table 8
Results of Propensity-score matching.
Dependent Variables: SHORT-TERM DEBT LONG-TERM DEBT CASH HOLDINGS CASH DIVIDENDS
(1) (2) (3) (4)
Coefficient Coefficient Coefficient Coefficient
t-Statistic t-Statistic t-Statistic t-Statistic

EXTREME RAINFALL − 0.210*** 0.168*** 0.236*** − 0.441***

(− 3.01) (5.03) (3.59) (− 5.33)
SIZE 0.011*** 0.012*** 0.005*** 0.025***
(7.52) (12.82) (2.69) (10.88)
LNAGE 0.011*** 0.003*** − 0.024*** − 0.033***
(5.82) (2.61) (− 11.06) (− 12.52)
INTANGIBLE 0.129*** − 0.000 − 0.302*** − 0.126**
(2.75) (− 0.01) (− 6.33) (− 2.41)
PPE 0.053*** 0.026** − 0.260*** − 0.171***
(3.50) (2.57) (− 17.65) (− 9.46)
SALES GROWTH 0.003 0.009*** 0.000 − 0.008***
(1.10) (4.43) (0.11) (− 2.65)
LNGDP − 0.005*** − 0.001 0.001 − 0.000
(− 2.71) (− 1.05) (0.68) (− 0.06)
GDP GROWTH 0.010 0.043* 0.178*** − 0.032
(0.26) (1.89) (4.22) (− 0.68)
LEVERAGE − 0.228*** − 0.071***
(− 21.10) (− 5.82)

Industry Fixed Effects Included Included Included Included

Year Fixed Effects Included Included Included Included
Adj. R-squared 0.196 0.369 0.432 0.154
Observations 6900 6900 6900 6900

Notes: This table presents the results of Propensity-score matching. See Appendix A for variable definitions. The t-statistics are clustered at the
company level. ***, **, and * denote statistical significance at the two-tailed 1%, 5%, and 10% levels, respectively.

682
S. Chen et al. Journal of Economic Behavior and Organization 216 (2023) 670–685

4.4.4. Other robustness tests

To avoid idiosyncratic observations perturbing the results of our main tests, we also eliminate some special samples. In China,
corporate headquarters may be different from their manufacturing and operation locations. Many firms headquartered in more
economically developed areas would set up factories and production bases in other areas for reasons including reducing costs,
convenient transportation and sales, and getting closer to raw materials origins. As a result, the precipitation of the meteorological
station closest to a firm’s headquarters cannot accurately represent the rainfall data faced by the firm. To avoid this interference, we
exclude firms with headquarters in Beijing, Shanghai, Guangzhou, and Shenzhen, as these firms generally set up production facilities in
other regions. Using the new sample, we re-regress our baseline model, and the results shown in Table 9 confirm that our main findings
remain robust.

5. Conclusion and discussion

Climate risk has generated significant socioeconomic impacts. Recently, uneven precipitation has heightened concerns regarding
weather-related anomalies, with the frequency and intensity of extreme rainfall becoming more severe. Extreme precipitation events
can lead to catastrophic losses for the nation and its economy. Given the limited research on the impact of extreme rainfall at the firm
level, this paper examines the correlation between extreme rainfall and corporate financial resilience. Intense precipitation events may
result in liquidity constraints and financial distress. Previous studies suggest that firms should enhance their organizational resilience
during extreme weather events (e.g., Berkhout et al., 2006; Linnenluecke and Griffiths, 2010; Huang et al., 2017). Consistent with
these findings, based on Chinese data, we discover that extreme rainfall can influence firms’ financing policies. Specifically, an in­
crease in extreme precipitation events leads to a decrease in short-term debts, an increase in long-term debts, an augmentation of cash
reserves, and a reduction in cash dividend distribution. These results demonstrate that firms tend to maintain financial slack resources
under high environmental risks. Additionally, we also provide evidence of the adverse effect of extreme rainfall on firms’ operational
activities and operating uncertainty. Furthermore, cross-sectional tests reveal that the relationship between extreme rainfall and firms’
financing policies is stronger in sectors sensitive to rainfall but weaker for state-owned companies or those located in cities with a
higher number of financial institutions or geographically dispersed factories. Our findings remain robust after conducting several
sensitivity analyses such as controlling for firm fixed effects using the instrumental variable approach and PSM method while
excluding special samples.
Our research examines how firms address environmental risks and climate challenges, highlighting the importance of organiza­
tional resilience as a crucial strategy for mitigating potential losses caused by extreme weather events. As extreme rainfall becomes
more frequent, firms must allocate efficient resources to deal with uncertainties and financial distress following unexpected events.
The adaptability and flexibility of a firm’s financial structure can indicate its ability to cope with climate risks, leading us to
demonstrate that firms will adjust their financial policies in response to extreme rainfall. This provides new insights into understanding
the determinants of financial policies while also offering guidance for future research on means through which firms can manage

Table 9
Results of removing special samples.
Dependent Variables: SHORT-TERM DEBT LONG-TERM DEBT CASH HOLDINGS CASH DIVIDENDS
(1) (2) (3) (4)
Coefficient Coefficient Coefficient Coefficient
t-Statistic t-Statistic t-Statistic t-Statistic

EXTREME RAINFALL − 0.239*** 0.221*** 0.266*** − 0.354***

(− 3.69) (8.10) (4.73) (− 4.18)
SIZE 0.010*** 0.013*** 0.004** 0.027***
(6.21) (13.84) (2.25) (10.24)
LNAGE 0.013*** 0.004*** − 0.024*** − 0.034***
(6.90) (3.60) (− 12.71) (− 13.53)
INTANGIBLE 0.142*** − 0.009 − 0.267*** − 0.127**
(3.05) (− 0.31) (− 6.47) (− 2.36)
PPE 0.058*** 0.026*** − 0.246*** − 0.163***
(3.75) (2.96) (− 19.33) (− 8.60)
SALES GROWTH − 0.003 0.005*** − 0.001 − 0.008***
(− 1.16) (2.89) (− 0.57) (− 2.92)
LNGDP − 0.001 − 0.001 0.001 − 0.001
(− 0.29) (− 0.87) (0.33) (− 0.38)
GDP GROWTH − 0.035 0.046*** 0.101*** − 0.010
(− 1.14) (2.60) (3.21) (− 0.25)
LEVERAGE − 0.209*** − 0.079***
(− 20.75) (− 6.04)
Industry Fixed Effects Included Included Included Included
Year Fixed Effects Included Included Included Included
Adj. R-squared 0.210 0.345 0.422 0.155
Observations 14,010 14,010 14,010 14,010

Notes: This table presents the results by removing special samples. See Appendix A for variable definitions. The t-statistics are clustered at the
company level. ***, **, and * denote statistical significance at the two-tailed 1%, 5%, and 10% levels, respectively.

683
S. Chen et al. Journal of Economic Behavior and Organization 216 (2023) 670–685

climate risks. Ultimately, our findings have important implications for policymakers and businesses seeking to avoid harm caused by
external shocks.

Declaration of Competing Interest

All the authors declare that the work described was original research that has not been published previously, and not under
consideration for publication elsewhere, in whole or in part. There is no conflict of interest within our paper submission. And the
manuscript is approved by all authors for publication.

Data availability

Data will be made available on request.

Appendix A Variable definitions

This table describes the procedure used to compute each variable used in our analyses. Our data are obtained either through the
China Stock Market & Accounting Research (CSMAR) database, or the National Oceanic and Atmospheric Administration (NOAA)
dataset. All continuous variables are winsorized at 1% and 99% of the distribution.

Variables Definition

SHORT-TERM DEBT Short-term debt scaled by lagged assets

LONG-TERM DEBT Long-term debt scaled by lagged assets
CASH HOLDINGS Cash and short-term investment scaled by lagged assets
CASH DIVIDENDS Cash dividends scaled by lagged assets
ROA Net income divided by lagged total assets
ROA VOLATILITY The standard deviation of quarterly pre-tax income scaled by total assets over the preceding five fiscal years
OCF Operating cash flow divided by lagged total assets
OCF VOLATILITY The standard deviation of quarterly cash flows from operations scaled by total assets over the preceding five fiscal years
EXTREME RAINFALL The frequency of rainfall days within a calendar year scaled over 365
SIZE Natural logarithm of lagged total assets in RMB
LNAGE Natural logarithm of the difference between observed year and founding year
INTANGIBLE Intangible assets divided by lagged total assets
LEVERAGE Total debt divided by total assets
PPE Net property, plant, and equipment divided by lagged total assets.
SALES GROWTH Sales change computed scaled by sales in the last fiscal year
LNGDP Natural logarithm of GDP per capita
GDP GROWTH Annual growth of total GDP
RAIN-SENSITIVE An indicator that equals one for rain-sensitive industries and zero otherwise. Following Rao et al. (2021), we define rainfall-sensitive
INDUSTRIES sectors as agricultural machinery, agriculture & processed food, air transport services, tourism, hotels & restaurants, auto sector,
construction & allied activities, courier services, transport services, electricity generation & transmission, fertilizers & pesticides, and
mining & quarrying.
WIDELY DISTRIBUTED The average precipitation of cities where different factories of a firm are located.
STATE-OWNED An indicator that equals one for state-owned enterprises, and zero otherwise.
PROPERTY
# FINANCIAL Natural logarithm of the number of financial institutions in a city
INSTITUTIONS
SLOPE This variable is computed by first dividing each city’s administrative area into ten square km grids based on longitude and latitude
data. Next, calculate the difference between the maximum and minimum elevations for each grid. The SLOPE for each city is the
average of these differences.
HIGH TEMPERATURE The annual occurrence of days with temperatures exceeding a threshold of 30◦ C, expressed as a ratio to the total number of days in a
calendar year.

References

Addoum, J.M., Ng, D.T., Ortiz-Bobea, A., 2019. Temperature shocks and industry earnings news. SSRN Electron. J. https://doi.org/10.2139/ssrn.3480695.
Axelson, U., Jenkinson, T., Stromberg, P., Weisbach, M.S., 2013. Borrow cheap, buy high? the determinants of leverage and pricing in buyouts. J. Financ. 68 (6),
2223–2267.
Bai, J., Chu, Y., Shen, C., Wan, C., 2020. Managing climate change risks: sea level rise and mergers and acquisitions. SSRN Electron. J. https://doi.org/10.2139/
ssrn.3739599.
Bajaj, Y., Kashiramka, S., Singh, S., 2021. Application of capital structure theories: a systematic review. J. Adv. Manag. Res. 18 (2), 173–199.
Bates, T.W., Kahle, K.M., Stulz, R.M., 2009. Why do U.S. firms hold so much more cash than they used to? J. Financ. 64, 1985–2021.
Ben-Nasr, H., Boubaker, S., Rouatbi, W., 2015. Ownership structure, control contestability, and corporate debt maturity. J. Corp. Financ. 35, 265–285.

684
S. Chen et al. Journal of Economic Behavior and Organization 216 (2023) 670–685

Berkes, F., Colding, J., Folke, C., 2003. Navigating Social–Ecological systems: Building resilience For Complexity and Change. Cambridge University Press, Cambridge.
Berkhout, F., Hertin, J., Gann, D.M., 2006. Learning to adapt: organizational adaptation to climate change impacts. Clim. Change 78 (1), 135–156.
Bernstein, A., Gustafson, M.T., Lewis, R., 2019. Disaster on the horizon: the price effect of sea level rise. J. Financ. Econ. 134 (2), 253–272.
Chan, K.C., Chen, T., Liu, B., Wu, J., 2022. Air pollution and CEO compensation: evidence from China. J. Econ. Manag. Strategy 31, 448–469.
Choi, D., Gao, Z., Jiang, W., 2020. Attention to global warming. Rev. Financ. Stud. 33 (3), 1112–1145.
Cronqvist, H., Makhija, A.K., Yonker, S.E., 2012. Behavioral consistency in corporate finance: CEO personal and corporate leverage. J. Financ. Econ. 103 (1), 20–40.
Cvijanovic, D., van de Minne, A., 2021. Does Climate Change Affect Investment performance: Evidence from Commercial Real Estate. MIT Center for Real Estate
Research Paper.
Dehejia, R.H., Wahba, S., 2002. Propensity score matching methods for nonexperimental causal studies. Rev. Econ. Stat. 84 (1), 151–161.
Dell, M., Jones, B., Olken, B.A., 2014. What do we learn from the weather? the new climate-economy literature. J. Econ. Lit. 52 (3), 740–798.
Duchin, R., 2010. Cash holdings and corporate diversification. J. Financ. 65 (3), 955–992.
Easterbrook, F.H, 1984. Two agency-cost explanations of dividends. Am. Econ. Rev. 74 (4), 650–659.
Filippova, O., Noy, I., Nguyen, C., Rehm, M., 2019. Who cares? Future sea-level-rise and house prices. CESifo Work. Pap. https://doi.org/10.2139/ssrn.3374587.
Fischer, E.O., Heinkel, R., Zechner, J., 1989. Dynamic capital structure choice: theory and tests. J. Financ. 44 (1), 19–40.
Fu, X., Jan, N., 2021. Sea level rise, homeownership, and residential real estate markets in south Florida. Prof. Geogr. 73 (1), 62–71.
Fuerst, F., Warren-Myers, G., 2019. Sea level rise and house price capitalization. SSRN Electron. J. https://doi.org/10.2139/ssrn.3359289.
Gao, H., Harford, J., Li, K., 2013. Determinants of corporate cash policy: insights from private firms. J Financ. Econ. 109 (3), 623–639.
Goldsmith-Pinkham, P.S., Gustafson, M., Lewis, R., Schwert, M., 2021. Sea level rise exposure and municipal bond yields. SSRN Electron. J. https://doi.org/10.2139/
ssrn.3478364.
Hoffmann, V.H., Sprengel, D.C., Ziegler, A., Kolb, M., Abegg, B., 2009. Determinants of corporate adaptation to climate change in winter tourism: an econometric
analysis. Glob. Environ. Chang. 19 (2), 256–264.
Hong, H., Li, F.W., Xu, J., 2016. Climate risks and market efficiency. SSRN Electron. J. https://doi.org/10.2139/ssrn.2776962.
Huang, H.H., Kerstein, J., Wang, C., 2017. The impact of climate risk on firm performance and financing choices: an international comparison. J. Int. Bus. Stud. 49 (5),
633–656.
Jensen, M., 1986. Agency costs of free cash flow, corporate finance, and takeovers. Am. Econ. Rev. 76, 797–825.
Jiang, F., Li, C.W., Qian, Y., 2019. Can firms run away from climate-change risk? Evidence from the pricing of bank loans. SSRN Electron. J. https://doi.org/10.2139/
ssrn.3477450.
Jin, X., Liang, S., Yu, J., 2022. Management geographical proximity and stock price crash risk. China Financ. Rev. Int. 12 (4), 601–622.
Keys B.J., and P. Mulder. 2020. Neglected no more: housing markets, mortgage lending, and sea level rise. NBER Working Paper. 10.3386/W27930.
Kotz, M., Levermann, A., Wenz, L., 2022. The effect of rainfall changes on economic production. Nature 601 (7892), 223–227.
Kraus, A., Litzenberger, R.H., 1973. A state-preference model of optimal financial leverage. J. Financ. 28 (4), 911–922.
Kumar, A., Xin, W., Zhang, C., 2019. Climate sensitivity and predictable returns. SSRN Electron. J. https://doi.org/10.2139/ssrn.3331872.
Liang, X.Z., 2022. Extreme rainfall slows the global economy. Nature 601 (7892), 193–194.
Lin, C., Schmid, T., Weisbach, M.S., 2019. Climate change, operating flexibility and corporate investment decisions. SSRN Electron. J. https://doi.org/10.2139/
ssrn.3478625.
Lin, X., Li, A., Zhang, P., Chen, W., 2022. The disciplinary role of product market competition on cash holding. Int. Rev. Econ. Financ. 83, 653–671.
Linnenluecke, M.K., Griffiths, A., Winn, M.I., 2008. Organizational adaptation and resilience to extreme weather events. Paper Presented At the Annual Meetings of
the Academy of Management. (Recipient of the Carolyn Dexter Best International Paper Award), Anaheim, CA.
Linnenluecke, M.K., Griffiths, A., 2010. Beyond adaption: resilience for business in light of climate change and weather extremes. Bus. Soc. 49 (3), 477–511.
Liu, S., Weng, R., Yang, D., 2017. Natural disaster, fiscal pressure and tax avoidance: a typhoon-based study. China J. Account. Stud. 5 (4), 468–509.
Lotfaliei, B., 2018. Zero leverage and the value in waiting to have debt. J. Bank. Financ. 97 (12), 335–349.
Marobhe, M.I., Kansheba, J.M.P., 2022. Stock market reactions to COVID-19 shocks: do financial market interventions walk the talk? China Financ. Rev. Int. 12 (4),
623–645.
Marshall, A.P.S.R, Roy, P.P., Thapa, C., 2021. Extreme rainfall and institutional investor behavior. SSRN Electron. J. https://doi.org/10.2139/ssrn.3815526.
Mclean, R.D., 2011. Share issuance and cash savings. J. Financ. Econ. 99 (3), 1697–1727.
Murfin, J., Spiegel, M., 2020. Is the risk of sea level rise capitalized in residential real estate? Rev. Financ. Stud. 33 (3), 1217–1255.
Nguyen, T., Bai, M., Hou, G., Truong, C., 2022. Drought risk and capital structure dynamics. Account. Financ. 62 (3), 3397–3439.
Opler, T., Pinkowitz, L., Stulz, R., Williamson, R., 1999. The determinants and implications of corporate cash holdings. J. Financ. Econ. 52 (1), 3–46.
Pan, Y., Shroff, N., Zhang, P., 2023. The dark side of audit market competition. J. Account. Econ. 75 (1), 101520.
Pankratz, N., Bauer, R., Derwall, J., 2019. Climate change, firm performance, and investor surprises. SSRN Electron. J. https://doi.org/10.2139/ssrn.3443146.
Pankratz, N., Schiller, C., 2021. Climate change and adaptation in global supply-chain networks[C]//Proceedings of Paris December 2019 Finance Meeting
EUROFIDAI-ESSEC, European Corporate Governance Institute–Finance Working Paper (775).
Park, J., Pankratz, N.M.C., Behrer, A., 2021. Temperature, workplace safety, and labor market inequality. SSRN Electron. J. https://doi.org/10.2139/ssrn.3892588.
Pinkowitz, L., Stulz, R., Rohan, W., 2006. Does the contribution of corporate cash holding and dividends to firm value depend on governance? a cross-country a
nalysis. J. Financ. 61, 2725–2751.
Rao, S., Koirala, S., Thapa, C., Neupane, S., 2021. When rain matters! Investments and value relevance. J. Corp. Financ. 73, 101827.
Rubin, D.B., 1973. Matching to remove bias in observational studies. Biometrics 29, 159–183.
Titman, S., 1984. The effect of capital structure on a firm’s liquidation decision. J. Financ. Econ. 13 (1), 137–151.
Tschakert, P., Dietrich, K., 2010. Anticipatory learning for climate change adaptation and resilience. Ecol. Soc. 15 (2), 11.
Vogus, T., Sutcliffe, K., 2007. Organizational resilience: towards a theory and research agenda. In: Proceedings of the IEEE International Conference on Systems, Man
and Cybernetics. Montreal, Quebec, pp. 3418–3422.
White, B.A., Jakob, C., Reeder, M.J., 2022. Fundamental ingredients of Australian rainfall extremes[J]. J. Geophys. Res.: Atmospheres 127 (17), e2021JD036076.
Woods, D., 2006. Essential characteristics of resilience. Resil. Eng. Concepts Precepts 21–34.
Zhang, J., Chen, H., Zhang, P., Jiang, M., 2022. Product market competition and the value of corporate cash: an agency theory explanation. Int. Rev. Financ. Anal. 84,
102422.

685