Introduction

Climate change disrupts the ecosystem and causes unprecedented damage to economic development and human health (Jung et al., 2018). The United Nations Framework Convention on Climate Change, the Kyoto Protocol, and the Paris Agreement have been successfully adopted and implemented to cope with climate change. The Chinese government ratified the United Nations Framework Convention on Climate Change in November 1992. In August 2002, the Chinese government approved the Kyoto Protocol. In September 2016, the Chinese government ratified the Paris Agreement. To implement these agreements, the Chinese government has revised the Constitution, the Environmental Protection Law, the Coal Law, and the Electricity Law and formulated laws such as the Energy Conservation Law, the Renewable Energy Law, the Air Pollution Prevention Law, the Cleaner Production Promotion Law, the Circular Economy Promotion Law, and the Measures for the Administration of Certification of Energy Conservation and Emission Reduction Products. The Chinese government has also proposed a 1 + N policy system to achieve the goals of peak carbon and carbon neutrality. 1 is the Opinions of the Central Committee of the Communist Party of China and the State Council on Fully, Accurately and Comprehensively Implementing the New Development Concept and Doing a Good Job in Carbon Emission Peak and Carbon Neutralization, which plays a leading role in the "1 + N" policy system of peak carbon and carbon neutralization. N is the Peak Carbon Action Plan Before 2030, which is responsible for building a peak carbon and carbon neutrality policy system with clear objectives, a reasonable division of labor, strong measures, and an orderly connection.

All enterprises inevitably generate carbon emissions in the process of production and operation (Subramaniam et al., 2015). Therefore, minimizing carbon emissions while maximizing economic output is necessary to achieve peak carbon emissions and carbon neutrality. At the macro level, this means that the more GDP generated per unit of carbon emissions, the more peak carbon and carbon neutrality may be realized. In this study, the GDP per unit of carbon emissions in a region is regarded as the regional carbon efficiency. The higher the regional carbon efficiency is, the higher the GDP per unit of carbon emissions. Achieving such a policy makes it more likely for the regional government to strike a balance between economic development and carbon emissions reduction, reducing the pressure on it to achieve peak carbon and carbon neutrality. Thus, the carbon risk imposed by the government on enterprises is smaller.

At the same time, a sharp rise in corporate cash holdings has become a global phenomenon in the past 30 years, and the cash holdings of listed companies in the United States have continued to rise (Bates et al., 2009; Gao et al., 2013). High cash holdings in Chinese enterprises have become increasingly prominent, leading to diseconomies of scale. Cash is a low-income resource for enterprises, and holding too much cash is not conducive to the development of enterprises. An enterprise’s cash over-holdings lead to more costs (Lian and Su, 2008). First, there is an opportunity cost of too much cash holdings for internal financing instead of using external funding with a lower capital cost. Second, the management cost of corporate cash holdings. Third, too much cash holdings will bring more agency costs, especially in countries with poor corporate governance mechanisms, such as China. Cash is more likely to be used by managers for personal gain (Myers and Rajan, 1998).

Based on the above research background, this study aims to investigate the impact of regional carbon efficiency on an enterprise’s cash holdings and its mechanism. Corporate cash holdings are closely related to the prevention motivation of managers (Bates et al., 2009; Faulkender and Wang, 2006; Riddick and Whited, 2009; Keynes, 1936). Uncertainty about the government’s carbon reduction policy induces carbon risk, increasing uncertainty about enterprise cash flow and enhancing the financial constraints on enterprises. Hence, improvement in regional carbon efficiency can reduce carbon risk, which can help improve the stability of enterprise cash flow, alleviate enterprise financing constraints, and weaken the firm’s prevention motivation. Given this, the qualitative analysis of this study posits that regional carbon efficiency can reduce corporate cash holdings. In addition, as a cash destination, the demand for carbon efficiency may impel enterprises to make long-term investments, ultimately reducing cash holdings. Therefore, the qualitative analysis of this study contends that regional carbon efficiency affects the cash holdings of enterprises through their debt financing capability and long-term investment level.

Furthermore, the study empirically tests the hypotheses proposed by the qualitative analysis. These are tested empirically by constructing a two-way fixed effects model of enterprises and years and a mediating effect test model, respectively. Thus, this study has several important findings. First, the higher the regional carbon efficiency is, the lower the cash holdings level of the local enterprises. This result remains valid after robustness testing as well. Second, based on reducing the carbon risks enterprises face, regional carbon efficiency can improve the debt financing capability of enterprises, encourage enterprises to make long-term investments, and reduce corporate cash holdings. Our conclusion remains when we replace the independent variables and exclude endogenous factors.

Our research contributes to several streams of literature. First, we contribute to the growing literature on corporate cash holdings. Few studies explore how regional carbon emissions affect corporate cash holdings. This study discusses the impact of regional carbon emissions on corporate cash holdings, enriching the literature on the factors influencing cash holdings. Second, our research contributes to the emerging literature on the economic consequences of climate change and regional carbon efficiency. The literature lacks a discussion on the impact of regional carbon efficiency on micro-enterprises. This study enriches the literature on climate change and the economic consequences of regional carbon efficiency from the perspective of corporate cash holdings.

The remainder of this study is arranged as follows. Section 2 provides a literature review. Section 3 describes our empirical design. Section 4 reports our empirical results and robustness tests. Our discussion is found in Section 5. Finally, we provide a summary.

Literature review

Factors influencing corporate cash holdings

With the change in business objectives, cash holding management has become the center of modern enterprise financial management (Ni and Sun, 2021). Cash holdings can increase the value of enterprises (Chan et al., 2022; Liu et al., 2022). Many other factors influence the cash holdings of the enterprise. Keynes, in his 1936 General Theory of Employment, Interest, and Money, took the lead in studying the problem of corporate cash holdings and proposed for the first time three major motivations of corporate cash holdings: precautionary, transaction, and investment motivations. Much research on precautionary motivation has recently been particularly active (Honda, 2023; Magerakis and Habib, 2022; Park, 2022; Yuan and Gao, 2022; Yeh et al., 2022; Zhang and Zhou, 2022). Credit lines represent an important alternative to cash as a source of liquidity (Honda, 2023). However, the cash holdings of enterprises with credit lines are lower than those without (Honda, 2023). Enterprise innovation increases with market competition, and cash holdings of enterprises increase with enterprise innovation as well; the financial constraints of enterprises increase with market competition, and cash holdings of enterprises decrease with financial constraints (Zhang and Zhou, 2022). The digital economy develops corporate low-carbon innovation by alleviating corporate financing constraints and environmental uncertainty (Chen, 2023), which can influence corporate cash holdings. Enterprises in more corrupt countries hold cash beyond their optimum for the given cost of carry due to severe financial constraints (Park, 2022). Based on the precautionary motive for cash holdings, enterprises will hold more cash when operating in an environment of high uncertainty (Magerakis and Habib, 2022). There is a significant positive effect of green credit policy on the cash holdings of enterprises because green credit policy reduces the bank loans and liabilities of enterprises, exacerbating the external financing environment (Yuan and Gao, 2022). Moreover, economic and political uncertainty and good corporate governance affect the cash holdings of enterprises (Cui et al., 2022).

Economic changes in regional carbon efficiency

Another branch of literature related to this study is research on the economic changes in regional carbon efficiency. As the macro environment of corporate production and operation, regional carbon efficiency affects the uncertainty of local government carbon reduction regulations. The lower the regional carbon efficiency is, the higher the uncertainty about the government’s implementation of carbon reduction regulations. Uncertainty about government carbon reduction regulations will induce carbon risks (Lin and Wu, 2022), including regulatory, physical, and business risks (Jung et al., 2018), generating adverse effects for enterprises (Chen, 2022). Enterprises will hold cash to avoid carbon risks as a cushion in the face of uncertainty. The adverse effects of carbon risks on enterprises are as follows. First, carbon risks bring direct costs to enterprises. For example, carbon risk will make the government consider imposing taxes to punish enterprises (Wang, 2020), causing enterprises additional potential litigation and compliance costs (Zhou et al., 2017). Second, policy uncertainty affects the leverage ratio. Affected by carbon risk, corporate profits decrease with an increase in risk, net assets decrease, and the leverage ratio increases (Hsu et al., 2023). Carbon risk will lead to higher financial distress risk, forcing enterprises to actively reduce leverage (Nguyen and Phan, 2020). Third, uncertainty leads to increased financing constraints. Creditors will reduce the impact of the borrower’s carbon risk by controlling the loan contract terms of collateral, debt maturity, and debt price (Jung et al., 2018), increasing the cost of corporate debt financing (Jung et al., 2018; Wang, 2020), and enhancing the financing constraints on enterprises. In addition, carbon risk urges enterprises to reduce acquisitions of domestic enterprises but increase acquisitions of foreign enterprises (Bose et al., 2021), significantly affecting the credit risk of enterprises (Dumrose and Höck, 2023).

Economic impact of climate change

The climate is also one of the macro environments for corporate production and operation. Regional carbon efficiency has an impact on local climate change. Therefore, the stream of research on the economic impact of climate change is also related to this study. Climate change induces global temperature rise, with attendant impacts on global and local economies, which has long been the subject of scholarly research. Many studies focus on macroeconomic impacts (e.g. Chen and Yang, 2019; Hansen, 2022; Kumar and Khanna, 2019). Climate change has a long-term negative impact on economic growth (Kahn et al., 2021), poses an important policy challenge for central banking (Hansen, 2022), temperature rise reduces economic output (Magazzino et al., 2021), and extreme temperature levels hinder economic productivity (Kumar and Khanna, 2019; Magazzino et al., 2021). Another stream of this research focuses on microeconomic impacts (Li et al., 2018; Yu et al., 2019; Zhang et al., 2018). For example, temperature has a nonlinear effect on the productivity of enterprises, and increases in summer temperature significantly negatively affect enterprises’ productivity. Urban households will adaptively increase energy consumption at low temperatures and buy more air conditioners in high-temperature weather (Yu et al., 2019). Therefore, climate warming in the summer in China will cause households to consume more energy than climate cooling in winter (Li et al., 2018). Moreover, Banks seem to be aware of the consequences of climate change on their business, but they are still very timid regarding operational implementation (Caby et al., 2022).

Generally, scholars have researched the effect of carbon risks on cash holdings, climate change, and regional carbon efficiency. However, scholars have seldom explored the relationship between regional carbon emissions and corporate cash holdings.

Methods

Variables

Corporate cash holdings level

The dependent variable of this study is the corporate cash holdings level (Cash1), following previous studies (Bates et al., 2009; Chen and Lu, 2019; Xiong et al., 2020). This study obtains Cash1 from cash and cash equivalents/total assets as the proxy variable of corporate cash holdings level.

In addition, Chris and Sushil (2018) regard cash and trading financial assets as cash. Therefore, referring to the literature (Xiong et al., 2020; Yang and Yin, 2018), this study also takes (cash + trading financial assets)/(total assets - cash and cash equivalents) to obtain Cash2 as a proxy variable for the cash holdings level.

In addition, several studies have also removed the cash holdings level from the industry average to eliminate the industry impact (Zhou et al., 2020). Here, we subtracted the industry average from Cash1 and Cash2 to obtain rCash1 and rCash2 for the robustness test.

Regional carbon efficiency

The independent variable in this study is regional carbon efficiency. Here, the regional carbon efficiency (Ceff) is characterized by the regional carbon efficiency index. When Wang and Xu (2015) studied the decoupling of haze, the haze efficiency was represented by the GDP of each region divided by the amount of haze. This study defines regional carbon efficiency as the economic output in exchange for regional unit carbon emissions. Therefore, based on Wang and Xu (2015), the regional carbon efficiency index Ceff is obtained by dividing the actual GDP of each prefecture-level city by the carbon emissions of prefecture-level cities as the proxy variable of regional carbon efficiency. Meanwhile, rCeff is obtained by dividing the provincial actual GDP by the provincial carbon emissions for the robustness test.

Control variables

To overcome the effect of omitted variables as far as possible, following the literature (Bates et al., 2009; Chris and Sushil, 2018; Yang and Yin, 2018), this study select the followings important internal factors affecting corporate cash holdings as control variables: (1) TobinQ, the market value of equity at the end of the year plus the book value of debt divided by the total assets; (2) Enterprise Size is the natural logarithm of the total assets of the enterprise; (3) the age of the enterprise (Age) is the natural logarithm of the current year minus the year of establishment of the enterprise plus one; (4) shareholding ratio of the largest shareholder (First); (5) Growth refers to the annual growth rate of the enterprise’s operating revenue; (6) profitability (Roa), which is the profit margin of total assets; (7) asset turnover rate (Tatr), operating revenue divided by average total assets; (8) operating cash flow (Cf) is the net cash flow from operating activities divided by non-cash assets; (9) working capital (Nwc), net working capital divided by net assets; and (10) dividend and interest payment (Di) is the cash paid for distributing dividend profits and paying interest divided by total assets.

Models

Model for impact analysis

This study constructs a panel effects model to verify the impact of regional carbon efficiency on the level of corporate cash holdings. Additionally, to incorporate individual and time effects, following the literature (Bates et al., 2009; Chris and Sushil, 2018; Yang and Yin, 2018), we construct a two-way fixed effects model of enterprises and year:

$$Cash_{i,t} = \alpha _0 + \beta _1Ceff_{i,t} + \beta _2X + \alpha _i + \lambda _t + \varepsilon _{i,t}$$
(1)

Cashi,t is the cash holdings level of enterprise i in year t, Ceffi,t is the carbon efficiency of the region where enterprise i is located in the t-th year, and β1 is its coefficient. If β1 is significantly negative, an increase in the carbon efficiency of the region where the enterprise is located can reduce the cash holdings of the enterprise. X is the control variable, α0 is the intercept term, αi is the individual fixed effect of the i-th enterprise, λt is the fixed effect in the t-th year, and εi,t is the random error term.

Model for mechanism test

To unlock the black box of the mechanism of regional carbon efficiency on corporate cash holdings, this study refers to Chen and Zhu et al. (2022) and focuses on two channels that affect the financing situation of local enterprises, namely, "debt financing capacity" and "long-term investment". We believe that regional carbon efficiency reduces corporate cash holdings through enterprise debt financing capacity and long-term investment channels. To test the above transmission mechanism, we set the following model concerning the test procedure proposed by Wen and Ye (2014) and the approach of Chen and Yan et al. (2022).

$$Cash_{i,t} = \alpha _0 + \beta _1Ceff_{i,t} + \beta _2X + \alpha _i + \lambda _t + \varepsilon _{i,t}$$
(2)
$$m_{i,t} = \delta _0 + \phi _1Ceff_{i,t}\varphi Z + \delta _i + \lambda _t + \nu _{i,t}$$
(3)
$$Cash_{i,t} = \alpha _0 + \beta _1Ceff_{i,t} + \beta _2X + \delta m_{i,t} + \alpha _i + \lambda _t + \varepsilon _{i,t}$$
(4)

In the above model, mi,t are mediator variables, including debt financing capacity (Debt) and long-term investment level (Linv), as follows:

Debt financing capability (Debt): China’s financial system is dominated by banks (Wang and Wang, 2021; Xie and Kuang, 2020). Bank loans are the main source of enterprise financing (Chen et al., 2011; Xu and Chen, 2019). In addition, bill financing is a common approach to interest-bearing debt financing. Therefore, this study divides the year-end short-term loans, long-term loans, and notes payable by the total assets at the end of the year to represent debt financing capacity.

Long-term investment level (Linv): The enterprise’s long-term investment includes equipment investment and intangible assets. Therefore, based on Jiang et al. (2021), the "cash paid for the purchase of fixed assets, intangible assets, and other long-term assets" is divided by the total assets to represent the enterprise’s long-term investment level (Linv).

The test procedure is as follows. First, model formula (2) is estimated without adding mediator variables. If the coefficient β1 of the regional carbon efficiency index Ceffi,t is significant, it indicates that regional carbon efficiency has a total effect on the corporate cash holdings level. If so, we continued the follow-up analysis; otherwise, it was treated as a masking effect. Second, a regression is performed to model formula (3) to judge the impact of the regional carbon efficiency index on mediator variables. Third, we estimate model Eq. (4) by adding the intermediate variable. Suppose the coefficient ϕ1 of the regional carbon efficiency index Ceffi,t in model Eq. (3) and the coefficient δ of the intermediate variable mi,t in model Eq. (4) are significant. In that case, this indicates that an intermediate effect exists. At this time, if the coefficient β1 of the regional carbon efficiency index Ceffi,t in model Eq. (4) is significant, it indicates that the intermediate variable has a partial intermediate effect. It shows that it has a complete mediating effect if it is not significant. Fourth, if only one of ϕ1 in model equation (3) and δ in model Eq. (4) is significant, the mediation effect needs to be tested by the Sobel test.

In the model, Eqs. (2) and (4) are control variables, as in Eq. (1). In model Eq. (3) is the control variable, which depends on the mediator variable.

When considering debt financing capacity (Debt) as the mediator variable, referring to the practices of Chen and Zhang (2021) and Yang and Pang (2017), enterprise size (Size), profitability (Roa), growth, economic development level (lnGDP), and financial development level (Fsize) of the region where the enterprise is located are controlled, and foreign direct investment (Pfdi) in the region where the enterprise is located. Fsize is the loan balance of prefecture-level cities divided by the total regional GDP; Pfdi is the FDI of prefecture-level cities converted according to the foreign exchange rate of the current year divided by the total regional GDP, and lnGDP is the natural logarithm of the actual per capita GDP based on 2008 figures.

Taking the long-term investment level (Linv) as the mediator variable, following Feng and Yu (2019), enterprise size (Size), asset turnover (Tatr), the proportion of tangible assets (Tang), growth (Growth), enterprise age (Age), and operating cash flow (Cf) are controlled, where Tatr is the operating revenue divided by the average total assets, and Tang is the fixed assets divided by total assets.

Data

Sample

Since the global financial crisis in 2008, "deleveraging" has become an economic and financial issue of common concern worldwide. In addition, in 2008, the World Wide Fund for Nature or World Wildlife Fund launched the "China Low Carbon City Development Project," and Shanghai, along with Baoding in Hebei Province, became the first batch of pilot cities. In November 2009, the Executive Meeting of the State Council decided that by 2020, China’s carbon dioxide emissions per unit of GDP will be 40%–45% lower than in 2005. This study selects data from 2008 to 2019 for empirical testing. For the enterprise sample selection, we use all A-share listed companies as samples and process the samples as follows. (1) We eliminate all ST samples and *ST samples; (2) we eliminate all samples of financial enterprises; (3) we eliminate samples with missing data; (4) we exclude extreme samples where EBIT is higher than average total assets; and (5) we exclude extreme samples of working capital divided by net assets that are too high or too low (20 times). We also exclude samples with fewer than two observed values in the sample period. After the above processing, 26041 enterprise-year observations were obtained. In addition, to eliminate the influence of outliers, the continuous variables were winsorized tailed up and down by 1%.

Data source

Our data on the carbon emissions of cities and provinces are from the CEADS (China Emission Accounts and Datasets) database. With the joint support of several research institutions, including the British Research Council, Newton Foundation, National Natural Science Foundation of China, and Chinese Academy of Sciences, the database gathered scholars from multinational research institutions such as Britain, America, and central Europe to jointly compile China’s multi-scale carbon emission inventory. Based on the approach of Wu et al. (2021), this study calculates the carbon emissions of prefecture-level cities and provincial levels based on the carbon emissions at the county level. The "internal control index" was obtained from the DIB database of Shenzhen Dibo Enterprise Risk Management Technology Co., Ltd. The data on prefecture-level cities come from the China Urban Statistical Yearbook, while other data are taken from the National Bureau of Statistics and the Wind Database. The carbon emissions in 2018 and 2019 were linearly interpolated, and the interpolated data were removed for the robustness test.

Descriptive statistics

Table 1 presents the descriptive statistics of the main variables from 2008 to 2019. As shown in Table 1, the average cash holdings level, Cash1, expressed by dividing cash and cash equivalents by total assets, is 0.1709, the minimum value is 0.0008, and the maximum value is 0.9724, indicating that there is a large difference in the cash holdings levels of enterprises. Second, the average cash holdings level Cash2, represented by cash plus trading financial assets divided by non-cash assets, is 0.2995, the minimum value is 0.0005, and the maximum value is 35.5547. This also shows that there is a large difference in the cash holdings levels of enterprises. Third, debt financing capacity Debt and long-term investment Linv are the mediator variables required for the later mechanism test.

Table 1 Descriptive statistics of the main variables.
Full size table

Results

Results of impact analysis

Base regression

Taking Cash1 as the dependent variable, only independent variables are considered. The estimation results are shown in column (1) of Table 2. After adding the control variables listed above, the estimation results are shown in column (3) of Table 2. Following the same order and taking Cash2 as the dependent variable, the estimation results are given in columns (2) and (4) of Table 2, respectively.

Table 2 Base regression results of model Eq. (1).
Full size table

Table 2, columns (1) to (4) show that the coefficients of the independent variable regional carbon efficiency index (Ceff) are significantly negative at the 1% level. This indicates that the regional carbon efficiency of enterprises reduces the level of corporate cash holdings. The benchmark regression results confirm our hypothesis that regional carbon efficiency can reduce corporate cash holdings.

Robustness test

Adjusting the standard error in individual and time double clusters can overcome the impact of autocorrelation and heteroscedasticity on statistical inference (Petersen, 2009). Double clustering standard errors are used in columns (1) to (4) of Table 2 to increase the reliability of the estimation results. In columns (1) to (4) of Table 2, the independent variables are significant at the 1% level, which can be regarded as a robustness test. Here, we further test the robustness of our findings through endogenous processing, changing the measurement of dependent variables and independent variables, changing the estimation model, and eliminating the interpolated data.

Endogeneity treatment

As a micro variable, corporate cash holdings have difficulty affecting the macro variable of carbon efficiency of the regions where enterprises are headquartered, and the two-way causality between dependent and independent variables is difficult to establish. However, endogeneity problems may still occur because of measurement errors or unobservable factors in regional carbon efficiency.

Therefore, we follow Kim et al. (2014) and Chen and Yan et al. (2022) and take the mean value of the carbon efficiency index (ivCeff) of other prefecture-level cities in the same year as the instrumental variable. The reason is that it is difficult for the carbon efficiency indices of other prefecture-level cities to affect the cash holdings level of enterprises in this region, and ivCeff meets the requirements of exogeneity. At the same time, measurement errors and unobservable factors will affect the measurement of the carbon efficiency index of local cities. Therefore, other prefecture-level cities’ average carbon efficiency index is correlated with the local carbon efficiency index, and ivCeff meets the correlation requirements. Taking Cash1 as the dependent variable, the instrumental variable method (IV) was used to re-estimate model Eq. (1). The Cragg–Donald Wald F statistic is 16,000, which is far greater than the critical value of 16.38. ivCeff passes the weak instrumental variable test. The number of instrumental variables and endogenous variables is the same, and there is no need for an over-identification test. Therefore, ivCeff is a valid variable.

Taking Cash1 as the dependent variable, model Eq. (1) is re-estimated by IV, with the results given in column (1) of Table 3. Taking Cash2 as the dependent variable, model Eq. (1) is re-estimated by IV. The results are shown in column (2) of Table 3. In addition, there may be endogeneity between the control and dependent variables due to reverse causality. To alleviate the endogeneity problem of control variables, this study also lags the control variables by one period, takes Cash1 as the dependent variable, and uses IV to re-estimate the model Eq. (1). The results are shown in column (3) of Table 3. Column (4) of Table 3 shows the results with Cash2 as the dependent variable.

Table 3 Estimation results of instrumental variables in model Eq. (1).
Full size table

From columns (1) to (4) of Table 3, the Cragg–Donald Wald F statistics of the weak instrumental variable test are far greater than the critical value of 16.38 under a 10% error, meaning that ivCeff is a valid tool variable. The coefficients of the regional carbon efficiency index of the independent variables are significantly negative at the 1% or 10% level. After excluding endogeneity, our finding that regional carbon efficiency can reduce corporate cash holdings remains valid.

Changing the measurement of cash holdings level

Following the literature (Dou and Lu, 2016; Li et al., 2018; Yang and Yin, 2018; Zhou et al., 2020), we removed the industry average of Cash1 and Cash2 to obtain rCash1 and rCash2. Using them as dependent variables, model formula (1) is re-estimated, and the results are shown in columns (1) and (2) of Table 4, Panel A. In columns (1) and (2), the coefficients of the regional carbon efficiency index Ceff are significantly negative at the 1% level. Therefore, when we change the measurement of the cash holdings level, regional carbon efficiency still reduces corporate cash holdings.

Table 4 Robustness test and estimation results of model Eq. (1).
Full size table

Changing the measurement of the regional carbon efficiency index

Following the approach of Li et al. (2020) in testing the robustness of the provincial financial technology development level, this study takes the provincial actual GDP and carbon emissions to calculate the provincial carbon efficiency index (rCeff) as independent variables, uses Cash1 and Cash2 as dependent variables, and re-estimates the model formula (1). The results are shown in columns (3) and (4) of Table 4, Panel A. The estimation results show that the coefficients of the carbon efficiency index of the province where the enterprise is located are significantly negative at the 1% level. Therefore, when the measurement of the regional carbon efficiency index is changed, the conclusion that regional carbon efficiency can reduce corporate cash holdings remains robust.

Changing the estimation model

Previous studies (Chris and Sushil, 2018; Hanlon et al., 2017; Xiong et al., 2020) also estimated the cash holdings level by controlling for the fixed effect of industry and time. Unbalanced development is China’s basic national condition, and there are great differences in economic development levels among the regions where enterprises are located. Therefore, under the condition of controlling the fixed effects of industry, time, and region, we re-estimate model formula (1) with Cash1 and Cash2 as dependent variables. The results are shown in columns (1) and (2) of Panel B in Table 4. The estimation results also show that the conclusion that regional carbon efficiency can reduce corporate cash holdings is robust when the estimation model is changed.

Eliminate interpolated data

As mentioned above, when calculating the regional carbon efficiency index, the carbon emissions in 2018 and 2019 were linearly interpolated. Here, excluding the data of these two years, we take Cash1 and Cash2 as dependent variables and re-estimate the model Eq. (3). The results are shown in columns (3) and (4) of Panel B in Table 5. From the estimation results, the conclusion that regional carbon efficiency can reduce corporate cash holdings remains robust when the interpolated data are excluded.

Table 5 Estimated results of debt financing channels.
Full size table

In sum, excluding endogeneity, after changing the measurement of cash holdings level and regional carbon efficiency index, changing the estimation model, and excluding the interpolated data, we conclude that regional carbon efficiency reduces the cash holdings level of enterprises.

Mechanism test

Based on the above settings, the mechanism test is described below.

Debt financing channels

Taking debt financing capacity (Debt) as the mediator variable and Cash1 as the proxy variable of corporate cash holdings level, the Fe estimation model formulas (2) to (4) are adopted, and the results are shown in columns (1) to (3) of Panel An in Table 5. From Panel A of Table 5, the coefficient of the regional carbon efficiency index Ceff in column (1) is significantly negative at the 1% level. Regional carbon efficiency can reduce corporate cash holdings, and a total effect exists. The coefficient of Ceff in column (2) is significantly positive at the 1% level. An improvement in regional carbon efficiency can improve the debt financing capability of enterprises. In column (3), the coefficient of the mediator variable Debt is significantly negative at the 1% level, and debt financing capability significantly reduces the cash holdings level of enterprises, meaning that the mediator effect exists, while the coefficient of Ceff in column (3) is significantly negative at the 5% level. Therefore, debt financing plays a partial mediating role. Columns (2) and (3) show that regional carbon efficiency improves debt financing capacity, thus reducing the level of cash holdings. Therefore, debt financing channels exist.

When estimating model (3), enterprise debt financing capacity as a micro variable does not affect the variable of regional carbon efficiency, and it is difficult to establish a two-way causal relationship between debt financing capacity and regional carbon efficiency. However, similar to the base regression, the regional carbon efficiency index may still be endogenous because of strategic errors or unobservable factors. Here, ivCeff is taken as the instrumental variable, and IV is used to re-estimate the model Eq. (3). The results are shown in column (4) of Panel A in Table 5. In addition, cash can be used as a reserve fund for enterprises to repay debts, which plays a positive role in protecting the rights and interests of creditors. Therefore, the more cash held, the more creditors such as banks are willing to provide debt funds to enterprises. Thus, cash reserves can accelerate the speed of financing success (Chi and Xu, 2019). At the same time, cash holdings have a signaling effect on debt capacity (Natke and Falls, 2010), and cash as a liquid asset can increase debt capacity (Myers and Rajan, 1998). Thus, there is a two-way causal relationship between an enterprise’s cash holdings and its debt financing capacity. Hence, when debt financing capacity debt is used as the mediator variable to estimate model Eq. (4), Debt is endogenous. Here, following Kim et al. (2014) and Chen and Zhang (2021), we take the average value (ivDebt) of the debt financing capacity of other enterprises in the same year as the instrument variable. Together with ivCeff, the IV re-estimation model formula (4) is adopted. The results are shown in column (5) of Panel An in Table 5. Columns (4) and (5) of Panel A in Table 5 show that debt financing channels exist without endogeneity.

Taking Cash2 as the proxy variable of cash holdings level and re-estimating model Eqs. (2) to (4) in the same order that Cash1 is the cash holdings level yields the results in Panel B of Table 5. The results of Panel B also show that there are debt financing channels, and the conclusion remains valid without endogeneity. The finding that debt financing channels exist is thus robust.

In conclusion, the regional carbon efficiency level reduces corporate cash holdings through enterprise debt financing channels.

Long-term investment channels

Similar to the debt financing channel, taking the long-term investment level (Linv) as the mediator variable and Cash1 as the proxy variable of corporate cash holdings level, a Fe estimation model, Eqs. (2) to (4), is adopted. The results are listed in columns (1) to (3) of Panel An in Table 6. From the perspective of columns (1) to (3), the mediator effect of the long-term investment level exists and plays a partial mediating role. According to columns (2) and (3), regional carbon efficiency improves the long-term investment level, thus reducing the cash holdings level. Therefore, long-term investment channels exist.

Table 6 Estimation results of long-term investment channels.
Full size table

Similar to debt financing channels, when estimating model (3), it is difficult to establish a two-way causal relationship between the long-term investment level and regional carbon efficiency level. However, the regional carbon efficiency index may still be endogenous because of strategic errors or unobservable factors. Similar to the debt financing channel, taking ivCeff as the instrument variable and using IV to re-estimate model Eq. (3) results in column (4) of Panel A in Table 6. In addition, enterprises facing financing constraints and having low cash holdings levels are more inclined to postpone investment plans, delaying long-term investment. At the same time, as an important source of funds for long-term investment, cash has a soothing effect on long-term investment (Liu et al., 2015). Therefore, cash holdings also affect long-term investments. Thus, a two-way causal relationship exists between the enterprise’s cash holdings and long-term investment level. When model Eq. (4) is estimated with the long-term investment Linv as the mediator variable, Linv is endogenous. Here, following Kim et al. (2014) and Chen and Zhang (2021), taking the mean value of the long-term investment level (ivLinv) of other enterprises in the same year as the instrumental variable, together with ivCeff, we use IV to re-estimate the estimation model formula (4). The results are shown in column (5) of Panel A in Table 6. According to columns (4) and (5) of Panel A in Table 6, the conclusion remains valid when endogeneity is excluded.

Similarly, we consider Cash2 as the proxy variable of cash holdings level and re-estimate model Eqs. (2) to (4). The results are shown in Panel B of Table 6. The results of Panel B also confirm the long-term investment channels, and the conclusion remains valid without endogeneity.

In conclusion, the regional carbon efficiency level reduces corporate cash holdings through long-term investment channels.

Discussion

Our research finds that regional carbon efficiency can reduce the cash holdings of enterprises. This is consistent with the research conclusions of the literature. A low regional carbon efficiency increases uncertainty about the government’s implementation of carbon reduction regulations. This leads to fluctuations and uncertainty over the potential litigation costs and compliance costs of energy supply enterprises such as coal-fired power enterprises (Zhou et al., 2017), which will affect the continuity and price stability of the energy supply, increasing enterprise income uncertainty and cash flows (Jung et al., 2018; Oestreich and Tsiakas, 2015). Enterprises will face a high carbon risk if the regional carbon efficiency is low. Because of the impact of carbon risk, enterprise cash flow uncertainties increase, and the risk of enterprise debt default rises. To control credit risk, creditors such as banks may reduce the loan line or require enterprises to increase risk compensation, increasing the debt financing cost of enterprises (Jung et al., 2018; Kleimeier and Viehs, 2021) and enhancing the financing constraints of enterprises. Improvement in regional carbon efficiency will reduce the cash flow uncertainty of enterprises, easing their financing constraints and weakening the preventive motivation of enterprises to hold cash, thus reducing the cash holdings of enterprises.

We find that regional carbon efficiency reduces the cash holdings of enterprises by improving the debt financing level of enterprises. This is consistent with the conclusions of the literature. An increasing number of studies have found that creditors such as banks include the carbon risk faced by borrowers in their risk assessment process before credit approval and make credit decisions accordingly (Hoffmann and Busch, 2008; Herbohn et al., 2019). In this way, local enterprises face high carbon risk in areas with low-carbon efficiency, and creditors such as banks may refuse to lend, thus reducing the availability of debt financing. Conversely, an improvement in regional carbon efficiency will reduce the carbon risk faced by local enterprises, thus raising the probability of credit extension by creditors such as banks and improving the debt financing level of enterprises. This will reduce the prevention motivation of enterprises, thus reducing corporate cash holdings (Han and Qiu, 2007).

We also found that regional carbon efficiency reduced the cash holdings of enterprises by encouraging them to make long-term investments. This is consistent with the literature. According to the real options theory, under uncertain conditions, the delay and wait for real asset investment has an option value, and enterprises will be more cautious in investment, thus inhibiting irreversible real investment and reducing the long-term investment of enterprises (Leahy and Whited 1996; Wang et al., 2018). Improving regional carbon efficiency will reduce the uncertainty around local government carbon reduction regulations, thus reducing the option value of real asset investment, which impels enterprises to abandon their watch-and-wait strategies and actively make long-term investments such as real investment. A long-term investment is long-term, irreversible, and risky (Wang et al., 2018) and is a typical capital consumption behavior (Chen and Zhang, 2021). Cash is an important source of funds for the long-term investment of enterprises (Liu et al., 2015). Increased long-term investments by enterprises will consume their cash holdings, thus reducing the cash holdings of enterprises.

Conclusions

Based on the data of Chinese A-share listed companies from 2008 to 2019, this study aims to identify the impact and influencing mechanism of regional carbon efficiency on corporate cash holdings. The main results are as follows. First, with the improvement of regional carbon efficiency, the cash holdings level of enterprises can be reduced. The reason is that facing the dual tasks of reducing carbon emissions and promoting economic growth, the uncertainty of China’s local government’s carbon reduction policy will be more obvious. This is bound to have uncertain impacts on enterprises in many aspects, resulting in local enterprises facing carbon risk. Regional carbon efficiency will reduce carbon risk and weaken the preventive motivation of enterprises holding cash under the condition of stabilizing enterprise cash flow, alleviating financing constraints, and improving the availability of external funds. Second, regional carbon efficiency reduces the cash holdings of enterprises by improving corporate debt financing capability and promoting long-term investment. That is because improvements in regional carbon efficiency can reduce the carbon risks faced by enterprises. In terms of cash source, it can alleviate the financing constraints of enterprises, while in terms of cash destination, it may promote enterprises to make long-term investments and reduce cash holdings.

Our findings have both theoretical and practical significance. First, our findings can improve the initiative of Chinese enterprises to participate in peak carbon programs and carbon neutrality. Chinese enterprises’ active participation in peak carbon programs and carbon neutralization is critical to improving regional carbon efficiency. At the same time, improving regional carbon efficiency will ultimately reduce the cash holdings of enterprises and promote enterprise development. This will enhance the endogenous motivation of Chinese enterprises to participate in peak carbon programs and carbon neutrality. Second, our findings are of significant importance for governments in developing countries to achieve a better balance between ecological and economic goals. Reducing carbon emissions to address climate change and promoting enterprise development to stimulate economic growth are interrelated dilemmas developing countries face. Our findings show that from the perspective of cash holdings, reducing carbon emissions to address climate change and promoting enterprise development to stimulate economic growth can be balanced. Governments in developing countries can actively encourage a reduction in carbon emissions to achieve a win-win situation and improve economic growth.

This study also has some limitations. First, this study determines the regional carbon efficiency by GDP generated by unit carbon emissions. Although this method is simpler in terms of data acquisition and calculation, it also has some one-sidedness. Carbon dioxide is neither the only production input nor the only undesirable output factor for an enterprise. Moreover, this metric focuses only on economic outputs. Therefore, future studies can establish a more comprehensive carbon efficiency indicator system or measure total factor carbon efficiency. Second, this study focuses on the prevention motivation of cash holdings due to the carbon risk environment faced by enterprises. In reality, however, corporate cash holdings are influenced by various environmental factors, such as the internal governance environment, the external financial environment, and the policy environment. The internal governance environment can affect corporate cash holdings through agency motivation. In the weak development of traditional financial sector (banks and capital markets) regions, an enterprise will tend to hold more cash because of transaction and prevention. Government quality and tax policies also affect corporate cash holdings. Future research can analyze the heterogeneous impact of regional carbon efficiency on corporate cash holdings in different production and operational environments.