Air Pollution, Credit Ratings, and Corporate Credit Costs: Evidence from China

Abstract

From the perspective of credit ratings, this paper studies the impact of air pollution on corporate credit costs and the impact mechanism. Based on 2007–2022 data on A-share listed companies in the Chinese capital market, this paper uses a two-way fixed effects model to examine the impact of air pollution on corporate credit costs and the impact mechanism. The results show that air pollution increases the credit costs for enterprises because air pollution affects the sentiment of rating analysts, leading them to give more pessimistic credit ratings to enterprises located in areas with severe air pollution. The moderating effect analysis reveals that the effect of air pollution on the increase in corporate credit costs is more pronounced for high-polluting industries, manufacturing industries, and regions with weaker bank competition. Further analysis reveals that in the face of rising credit costs caused by air pollution, enterprises tend to adopt a combination strategy of increasing commercial credit financing and reducing the commercial credit supply to cope. Although this response behavior alleviates corporations’ own financial pressure, it may have a negative effect on supply chain stability. This paper provides new evidence that reveals that air pollution is an implicit cost in the capital market, enriching research in the fields of environmental governance and capital markets.

1. Introduction

Air pollution, as a byproduct of rapid economic development, has become one of the most pressing challenges faced by countries around the world. Due to its wide-reaching and often invisible nature, air pollution poses a greater threat to physical and mental health than many other forms of pollution [1,2]. In recent years, a growing body of literature has examined its impact on both individuals and economic systems. Studies have shown that air pollution increases the likelihood of hospitalization [3], reduces life expectancy [4], and affects various financial market behaviors, including investor sentiment [5], analyst forecasts [6], trading decisions [7], and corporate productivity [8].

While the physical consequences of air pollution are well documented, its psychological and behavioral impacts are receiving increasing attention. Research in environmental psychology suggests that exposure to pollution can induce emotional stress and pessimism among financial professionals [9], which may impair their decision-making ability. In corporate finance, credit cost plays a critical role in determining the efficiency of debt financing. Financial institutions evaluate credit risk and design debt contracts based on the credit ratings assigned to firms [10], which raises an important question. If air pollution influences the mood and sentiment of credit rating analysts, could it lead to more pessimistic ratings and, in turn, higher credit costs for firms?

Previous studies have explored the link between air pollution and corporate financing from different perspectives, including risk perception and government intervention. Some argue that pollution increases firm-level default risk and overall financial uncertainty, which results in higher financing costs [11,12]. Others suggest that pollution can prompt government support through green credit policies or subsidies, which help reduce financing constraints and encourage green investment [13,14]. These views, however, often overlook the role of individual behavior. In fact, the common mechanism underlying both market and policy reactions to pollution lies in its negative effect on human physical and mental health. This also applies to credit rating analysts, whose evaluations are central to the pricing of corporate credit risk.

Building on this behavioral perspective, this paper investigates whether air pollution affects corporate credit costs through its influence on credit rating behavior. Drawing on environmental psychology theory [9], we focus on the possibility that analysts’ emotional responses to pollution may influence their assessments of firm risk. This provides a new angle for understanding how environmental conditions, beyond traditional firm-level fundamentals, can shape financing outcomes in capital markets.

China presents an ideal context for this analysis. Over the past decade, the Chinese government has taken major steps to combat air pollution, including the 2013 launch of the Air Pollution Prevention and Control Action Plan and revisions to the Environmental Protection Law. According to the Ministry of Ecology and Environment, the national average PM2.5 concentration dropped from 68 micrograms per cubic meter in 2013 to 29 in 2022, a reduction of 57 percent. Despite this progress, China’s air quality remains below the World Health Organization’s recommended levels and continues to lag behind that of developed countries. These characteristics, combined with China’s representative economic structure and financial system, make it a suitable setting for examining the impact of air pollution on corporate credit costs. The findings of this study may also have reference value for other developing economies that prioritize economic growth while facing environmental challenges.

Based on this background, this paper uses annual average PM2.5 data from Washington University in St. Louis and panel data on Chinese A-share listed firms from 2007 to 2022. A two-way fixed effects model is employed to test the effect of air pollution on corporate credit costs. The results show that air pollution significantly increases firms’ credit costs, primarily by affecting the credit ratings issued by analysts. Further analysis reveals that the effect is more pronounced in high-polluting industries, manufacturing sectors, and regions with low banking competition. In response to increased credit costs, firms tend to rely more on commercial credit financing while reducing their supply of trade credit, which may adversely affect supply chain stability.

Compared with previous studies, the possible innovations of this paper include the following three aspects: (1) In practical terms, air pollution is the most challenging environmental problem faced by developing countries. Strengthening research on and the prevention of air pollution is highly important for optimizing the global ecological environment and achieving sustainable development. (2) In theoretical terms, this paper enriches the relevant literature on how environmental conditions affect market decision-making [6] and expands research on the impact of air pollution on people’s mental health [15]. This paper suggests that air pollution unrelated to financial markets can also affect corporate credit costs by altering the credit rating predictions of rating analysts, which is consistent with the conclusion of Li et al. (2020) [16]. (3) From the perspective of commercial credit, this paper broadens the understanding of the microeconomic impacts of air pollution. We find that to alleviate the rising credit costs caused by air pollution, enterprises usually adopt strategies of increasing commercial credit financing and reducing the commercial credit supply. These measures ease the financial pressure on enterprises but at the cost of sacrificing supply chain stability. This discovery provides a new perspective for understanding the trade-offs of enterprises in the process of environmental adaptation, and it has important implications for policy-makers in terms of balancing the relationship between air pollution control and the sustainable development of enterprises.

Figure 1 illustrates the research framework of this paper. The remaining sections are organized as follows: Section 2 presents the literature review, which primarily reviews studies on the economic consequences of air pollution and the factors influencing credit costs. Section 3 constructs a theoretical analysis mechanism, dissecting the specific paths through which air pollution affects corporate credit costs from the perspective of rating analysts’ credit rating processes for enterprises. Section 4 describes the data processing and model specifications. Section 5 tests the impact of air pollution on corporate credit costs and its underlying mechanism. Section 6 conducts an extended analysis, explores the heterogeneity of the relationship between air pollution and corporate credit costs, and examines corporate strategic choices in response to credit cost changes under environmental pollution from the dual dimensions of trade credit financing and supply. Section 7 summarizes the research conclusions, policy implications, and research limitations.

2. Literature Review

Air pollution has drawn increasing attention in the literature for its multifaceted impact on human health, firm behavior, and broader economic outcomes. A substantial body of research confirms that air pollution adversely affects both physical and mental health. It increases the risk of illness and hospitalization [3], causes physiological discomfort such as headaches and respiratory irritation through blood pressure fluctuations [17], and raises the incidence of cardiopulmonary diseases, thereby reducing life expectancy [4]. Beyond physical harm, air pollution is also linked to psychological stress, including increased depression, social withdrawal, and suicide risk [18,19]. These mental and physiological effects have been found to impair cognitive functioning, weakening decision-making ability and leading to suboptimal outcomes in personal and economic contexts [6,20].

Building on these findings, recent studies have turned to the economic consequences of air pollution at the organizational level. Firms operating in polluted regions face rising human resource costs, as employees increasingly consider environmental quality when choosing jobs [21]. In response, firms must offer higher compensation, which can crowd out strategic investment such as R&D [22]. Air pollution also affects specific business decisions: It reduces advertising efficiency [23], alters cash-holding strategies [24], and may disrupt logistics, leading to increased inventory and inefficiency [25]. While early studies view pollution regulation as a cost burden that depresses productivity [8,26], more recent work suggests that environmental quality may in fact support operational efficiency. In broader terms, air pollution imposes indirect social costs by raising healthcare expenditures and reducing labor productivity [27] while also exacerbating crime and social instability, thereby increasing public safety expenditures [28].

In parallel, extensive literature has examined what drives corporate credit costs. Research consistently shows that firms with higher earnings quality face lower credit costs due to reduced information asymmetry [29] and that strong corporate governance enhances transparency and investor confidence, particularly in emerging markets where legal protections are weaker [30]. Firms with better social responsibility performance also benefit from easier access to credit [31]. Likewise, transparent ESG disclosures improve firms’ standing with creditors [32], and high-quality financial reporting improves creditworthiness and lowers financing costs by allowing creditors to assess default risk more accurately [33].

However, while the literature on air pollution and on credit cost determinants is rich, these two streams have evolved largely in isolation. Most studies on air pollution have focused on its operational and social costs but have not considered its impact on firms’ financing conditions. Conversely, research on credit costs has concentrated on internal governance, reporting, and CSR-related factors but has largely overlooked external environmental conditions such as air pollution. Even more notably, few studies have investigated whether air pollution can affect credit costs through behavioral mechanisms—in particular, through the sentiment and judgment of credit rating analysts, who play a central role in the pricing of credit risk.

This paper seeks to bridge that gap by investigating whether and how air pollution affects corporate credit costs via analyst behavior. By drawing on insights from environmental psychology and behavioral finance and by incorporating observable measures of analyst sentiment—namely, rating downgrades and rating volatility—we aim to offer new evidence on the hidden cost of pollution in capital markets and contribute to the integration of environmental risk into corporate finance research.

3. Theoretical Framework

3.1. Air Pollution, Credit Ratings, and Corporate Credit Costs

As an important basis for creditors to judge the level of corporate credit risk, credit ratings can reduce the risk of information asymmetry [34] and affect the cost of corporate credit. A higher credit rating conveys information about the reliability and performance strength of an enterprise to the market [35], enabling the enterprise to obtain more credit financing at a lower cost [36]. Conversely, compared with a high credit rating, a low credit rating indicates that an enterprise has poor debt-servicing ability. As a result, financial institutions charge such an enterprise a higher credit default risk premium, thereby increasing corporate credit costs. Previous studies have shown that not only do the characteristics of credit rating agencies and corporate financial characteristics affect corporate credit ratings [37,38], but external operating environments such as the institutional environment and environmental risk can also lead rating analysts to adjust corporate credit ratings [39,40]. Therefore, as an important component of the external business environment of enterprises, air pollution may affect the credit ratings given by rating analysts.

According to environmental psychology theory, air quality, which is an important component of the environment, may have an impact on both the physiological and the psychological aspects of rating analysts. Credit rating work is essentially a highly complex cognitive activity that requires analysts to evaluate the credit information of borrowing companies while also considering multidimensional information such as macroeconomic trends, market dynamics, and industry characteristics. The rating results that they release are essentially a forward-looking prediction of the borrowing company’s future debt-paying ability, with significant time span characteristics. During this process, the subjective emotions of analysts undoubtedly become an undeniable factor affecting the accuracy of predictions. Specifically, scholars such as Kang et al. have confirmed that analysts often have low accuracy and are susceptible to subjective bias when faced with long-term forecasting tasks [41]. This phenomenon is further supported by Dong et al., who reported that long-term forecasting, due to its complexity, is more susceptible to the influence of analysts’ emotional fluctuations [6]. When air pollution intensifies, rating analysts are prone to negative emotions such as tension, anxiety, and depression. This emotional fluctuation may trigger analysts to excessively worry about their own health, which can then be reflected in their work attitude, manifested as apathy. Therefore, in environments with severe air pollution, rating analysts often tend to give more pessimistic credit ratings. Notably, the downgrading of credit ratings, which are important references for financial institutions to assess corporate credit risk, directly lead to the tightening of credit policies by financial institutions, manifested as the intensification of the “reluctance to lend” phenomenon, which in turn increases the loan interest rates of financial institutions for enterprises, ultimately increasing the financing costs for enterprises. Based on the analysis above, this paper proposes the following research hypothesis:

Hypothesis 1.

The intensification of air pollution leads to a significant increase in corporate credit costs.

3.2. Analysis of the Moderating Effect Based on High-Polluting Industries

As the main body of natural resource consumption and environmental pollution emissions, high-polluting enterprises face more uncertainty and risk in the eyes of rating analysts. When air pollution intensifies, the emotional response mechanism of rating analysts may be more significant, especially in high-polluting industries, where analysts may have more pessimistic expectations for their corporate environmental responsibility and long-term development capabilities. On the one hand, the large amount of exhaust gas, wastewater, and waste residue emitted by high-polluting enterprises during the production process increases their environmental compliance costs and may also lead to higher legal risks and regulatory pressures for these enterprises [42]. This makes it easy for analysts to excessively magnify the negative impact of environmental risk and further downgrade the credit rating of high-polluting enterprises. On the other hand, due to the heavy environmental burden of high-polluting enterprises, the market pays close attention to their future sustainable development and social responsibility [43]. Analysts may make overly pessimistic judgments on the basis of emotional reactions, especially in the context of worsening air pollution, which can lead to low credit ratings and thus increase the credit costs of enterprises. Therefore, this study expects that the promoting effect of air pollution on corporate credit costs is more pronounced for high-polluting industries than for low-polluting enterprises. Based on the analysis above, we propose the following:

Hypothesis 2.

The effect of air pollution on the increase in corporate credit costs is more pronounced for high-polluting industries.

3.3. Analysis of the Moderating Effect Based on Manufacturing Industries

As an important pillar of the real economy, the manufacturing industry has complex production processes, including raw material procurement, processing and manufacturing, and product assembly. Some processes, such as metal smelting and chemical production, are prone to generating large amounts of air pollutants. When air pollution intensifies, manufacturing enterprises may face stricter environmental regulations [44]. To meet environmental standards, enterprises need to increase their investment in environmental protection equipment and improve their production processes, which leads to an increase in production costs and affects the profitability and debt repayment ability of enterprises. When evaluating enterprises, rating analysts tend to focus on these cost changes and are more inclined to give pessimistic credit ratings. In addition, the production cycle of manufacturing products is relatively long, and there is a time delay between raw material input and product output and sales receipts. When air pollution is severe and analysts’ emotions are affected, analysts are more susceptible to the interference of subjective bias for long-term forecasting tasks such as analyzing manufacturing enterprises. They may underestimate the future debt-paying ability of manufacturing enterprises and subsequently downgrade their credit ratings because of concerns about the continued impact of air pollution on their long-term operations. Additionally, a credit rating downgrade will cause financial institutions to tighten their credit policies for manufacturing enterprises, for example, raising loan interest rates and increasing the financing costs for enterprises [35]. Hence, the positive impact of air pollution on the credit costs of manufacturing enterprises may be more prominent. Therefore, this paper proposes the following research hypothesis:

Hypothesis 3.

Compared with the credit costs of non-manufacturing industries, air pollution has a more significant effect on the credit costs of manufacturing enterprises.

3.4. Analysis of the Moderating Effect Based on the Level of Competition in the Regional Banking Industry

The degree of regional banking competition is a key external factor influencing the credit decisions of financial institutions. In regions with high banking competition, banks conduct more detailed and diversified credit assessments to capture customer resources [44]. When air pollution intensifies, rating analysts may issue pessimistic ratings due to its impact, but banks in highly competitive environments will cautiously balance risks and returns. If banks tighten credit and raise interest rates solely due to rating downgrades, they may lose high-quality customers. Therefore, banks do not make decisions based solely on credit ratings; rather, they also comprehensively consider factors such as corporate market share and industry status [45]. Even if the credit ratings assigned by analysts decline due to air pollution, banks may reduce risk premiums and offer flexible credit terms, thereby weakening the positive effect of air pollution on credit costs through ratings. In contrast, in regions with low competition, the number of banks is small, monopolies are strong, and there is a lack of competitive pressure. When air pollution leads rating analysts to be pessimistic, banks are more inclined to tighten credit and raise interest rates. Enterprises have weak bargaining power, forcing them to accept high credit costs [46]. On this basis, this paper proposes the following research hypothesis:

Hypothesis 4.

As the degree of banking competition decreases, the effect of air pollution on enhancing corporate credit costs is significantly strengthened.

4. Model and Data

4.1. Regression Model

In this paper, a two-way fixed effects model is used to examine the impact of air pollution on corporate credit costs:

$$Credit{c}_{i,t}=c+\beta Airp+{\displaystyle \sum _j{\alpha }^j{X}_{i,t}}+{\gamma }_t+{\theta }_i+{\epsilon }_{i,t}$$

(1)

In Equation (1), the dependent variable Creditc represents corporate credit costs, and the explanatory variable Airp represents the level of urban air pollution. X denotes the control variables, including 11 firm-level and regional-level variables: firm size (Size), financial leverage (Lev), growth ability (Growth), profitability (Roe), the cash flow level (Cash), board size (Board), the proportion of independent directors (Indp), listing age (Listage), the proportion of the largest shareholder (First), the economic development level (Mgdp), and the industrial structure (Indstr). γ represents the time fixed effects, θ represents the industry fixed effects, and ε is the random disturbance term.

4.2. Data Sources

This paper selects A-share listed companies in China’s capital market from 2007 to 2022 as the research sample. To avoid the adverse effects of abnormal samples, the initial sample is screened as follows: (1) Financial and insurance industries have significantly different business models from other industries, with particularities in financial characteristics and operational risks; thus, this paper excludes financial and insurance companies. (2) Observations of insolvent companies are excluded, as such firms face severe financial distress and their financial data may not reflect normal operating conditions, potentially interfering with the research results. (3) Listed companies with abnormal trading conditions (such as ST and *ST) are excluded, as these companies typically have abnormal financial conditions or other issues that may corrupt the results. (4) Observations with missing relevant data are excluded. After screening, a final sample of 33,524 firm-year observations is obtained. In addition, all continuous variables are winsorized at the 1% and 99% percentiles to exclude the influence of extreme values on the research conclusions. In terms of data sources, macroeconomic data are from the CElnet Statistics Database, and all other relevant data are from the CSMAR Database.

Table 1. Sample industry distribution.

Industry	Freq.	Percent
Agriculture, Forestry, Animal Husbandry, and Fisheries	444	1.32%
Mining	835	2.49%
Manufacturing	21,159	63.12%
Electricity, Heat, Gas, and Water Production and Supply Industry	1265	3.77%
Building Industry	971	2.90%
Wholesale and Retail Trade	1825	5.44%
Transportation, Warehousing, and Postal Services	1139	3.40%
Accommodation and Catering	107	0.32%
Information Transmission, Software, and Information Technology Services	2090	6.23%
Real Estate	1527	4.55%
Leasing and Business Services	421	1.26%
Scientific Research and Technical Services	386	1.15%
Water Conservancy, Environment, and Public Facilities Management	484	1.44%
Resident Services, Repairs, and Other Services	38	0.11%
Education	41	0.12%
Health and Social Work	79	0.24%
Culture, Sports, and Entertainment	352	1.05%
Comprehensive	361	1.08%
Total	33,524	100.00%

presents the industry distribution of the sample, which includes 33,524 firms across 18 industries, with a significant concentration in manufacturing, which accounts for 63.12% of the total. This reflects the central role of the manufacturing sector in both environmental impact and corporate credit dynamics. Other notable industries include Information Technology Services (6.23%), Wholesale and Retail Trade (5.44%), and Real Estate (4.55%). Sectors such as Electricity and Water Supply (3.77%), Transportation (3.40%), and Mining (2.49%) also have moderate representation. In contrast, industries like Accommodation and Catering (0.32%), Education (0.12%), and Resident Services (0.11%) are minimally represented. The distribution indicates a focus on pollution-intensive and capital-dependent sectors, which are most likely to be influenced by environmental regulations and credit market conditions. This industry structure provides a strong foundation for analyzing the relationship between air pollution, credit ratings, and corporate credit costs.

4.3. Variable Definitions

4.3.1. Dependent Variable (Creditc)

In accordance with the practices of Zhou et al. and Tan et al. [11,47], this paper measures corporate credit costs by multiplying the ratio of interest expenses to total corporate liabilities by 100. The larger the value is, the higher the credit cost.

4.3.2. Explanatory Variable (Airp)

In this work, air pollution (Airp) is measured by the average surface PM2.5 mass concentration in the prefecture-level cities where enterprises are located. The original data are sourced from the annual global surface PM2.5 concentration (μg/m³) dataset provided by Washington University in St. Louis, with a spatial resolution of 0.01° × 0.01°. This metric has significant advantages: PM2.5 has been designated the primary control indicator for air quality improvement in China [as stipulated in the “Air Quality Continuous Improvement Action Plan” issued by the State Council in November 2023], not only reflecting scientific validity but also being closely related to corporate behavioral decisions. Furthermore, the World Health Organization (WHO)–endorsed Air Quality Guidelines (AQGs) recognize PM2.5 as a critical component, underscoring its global recognition and authority in air quality assessment. A higher PM2.5 mass concentration indicates more severe air pollution.

4.3.3. Control Variables

In Model (1), variables that may affect credit costs are controlled for, namely, firm size (Size), financial leverage (Lev), growth ability (Growth), profitability (Roe), the cash flow level (Cash), board size (Board), the proportion of independent directors (Indp), listing age (Listage), the largest-shareholder shareholding ratio (First), the economic development level (Mgdp), and the industrial structure (Indstr). Year and industry fixed effects are also controlled for. The specific variable definitions are shown in

Table 2. Definitions of the main variables.

Variable Category	Variable Name	Variable Symbol	Variable Description
Dependent variable	Corporate credit costs	Creditc	(Interest expenses/Liabilities) × 100
Explanatory variable	Air pollution	Airp	Annual average surface PM2.5 mass concentration in cities
Control variables	Firm size	Size	Ln (Total assets of the company)
	Financial leverage	Lev	Total liabilities/Total assets
	Growth ability	Growth	Annual growth rate of the company’s operating income
	Profitability	Roe	Net profit/Net assets average balance
	Cash flow level	Cash	Net cash flows from operating activities/Total assets
	Board size	Board	Ln (Number of board members)
	Proportion of independent directors	Indp	Number of independent directors/Number of board members
	Listing age	Listage	Ln (YearObservation − YearEstablishment + 1)
	Largest shareholder shareholding ratio	First	Shareholding ratio of the largest shareholder/100
	Economic development level	Mgdp	Ln (per capita GDP)
	Industrial structure	Indstr	Value of the secondary industry/Regional GDP

.

5. Empirical Test Results and Analysis

5.1. Descriptive Statistics

Table 3. Descriptive statistics.

VarName	Mean	SD	Min	Median	Max
Creditc	1.693	1.429	0.007	1.317	6.226
Airp	4.545	3.275	1.535	3.793	22.353
Creditscore	2.9825	1.7935	0.000	4.000	5.000
Size	22.270	1.321	19.895	22.071	26.352
Lev	0.455	0.197	0.075	0.451	0.894
Growth	0.379	1.015	−0.745	0.128	7.059
Roe	0.067	0.190	−0.964	0.071	0.546
Cash	0.042	0.070	−0.170	0.043	0.240
Board	2.132	0.202	1.609	2.197	2.708
Indp	0.376	0.054	0.333	0.364	0.571
Listage	2.896	0.351	1.792	2.944	3.584
First	0.343	0.151	0.085	0.319	0.749
Mgdp	11.859	0.960	9.486	11.914	15.023
Indstr	0.405	0.114	0.159	0.414	0.653

reports the results of the descriptive statistical analysis for the main research variables. The mean value of Creditc is 1.693, with a standard deviation of 1.429. Notably, the standard deviation is close to the mean, and the minimum value (0.007) and maximum value (6.226) indicate significant differences in credit costs among the sample enterprises. Airp has a mean value of 4.545 and a standard deviation of 3.275, suggesting considerable variation in environmental pollution levels across regions; the minimum value of 1.535 and maximum value of 22.353 visually demonstrate the wide span of air pollution levels. Creditscore has a mean of 2.9825, a standard deviation of 1.7935, a median of 4, and ranges from 0 to 5, indicating that while some firms have low or no credit ratings, a substantial portion are clustered at higher rating levels. The mean value of Lev is 0.455, indicating that the overall debt level of the sample enterprises is moderate. Other variables also exhibit good statistical distributions and are not elaborated here.

5.2. Benchmark Regression Results

Table 4. Benchmark regression results.

Variables	(1)	(2)	(3)	(4)
	Creditc	Creditc	Creditc	Creditc
Airp	0.0418 ***	0.0174 ***	0.0138 ***	0.0183 ***
	(8.6814)	(3.8714)	(3.0479)	(3.9931)
Size			−0.0526 ***	−0.0468 ***
			(−3.9150)	(−3.4610)
Lev			1.1343 ***	1.1167 ***
			(13.4340)	(13.2962)
Growth			−0.1154 ***	−0.1125 ***
			(−10.7458)	(−10.5017)
Roe			−0.7026 ***	−0.7133 ***
			(−11.7717)	(−11.9710)
Cash			0.4519 ***	0.4124 ***
			(2.8838)	(2.6325)
Board			−0.1064	−0.1103
			(−1.2788)	(−1.3193)
Indp			0.0304	0.0442
			(0.1122)	(0.1647)
Listage			0.1552 ***	0.1412 ***
			(3.2989)	(3.0148)
First			−0.6260 ***	−0.5909 ***
			(−6.2441)	(−5.9540)
Mgdp				−0.0841 ***
				(−5.0432)
Indstr				0.3554 ***
				(2.5883)
Constant	1.5036 ***	1.6146 ***	2.3376 ***	3.0829 ***
	(52.6202)	(61.4892)	(6.8842)	(7.7749)
Year FE	No	Yes	Yes	Yes
Indus FE	No	Yes	Yes	Yes
Obs	33,524	33,524	33,524	33,524
Adj R2	0.0091	0.2746	0.3223	0.3261

Note: *** indicate significance at the 1% levels, respectively. The t values in parentheses are all adjusted for clustering at the individual level.

presents the benchmark regression results for the impact of air pollution on corporate credit costs. In Column (1), without including control variables or fixed effects, the regression coefficient of Creditc is 0.0418 and is positive and significant at the 1% level. In Column (2), after controlling for year and industry fixed effects, the regression coefficient of Creditc is 0.0174 and is still positive and significant at the 1% level. Column (3) adds firm-level control variables on the basis of Column (2), and the regression coefficient of Creditc is 0.0138 and is positive and significant at the 1% level. Column (4) further incorporates region-level control variables, and the regression coefficient of Creditc is 0.0183 and remains positive and significant at the 1% level. These results collectively indicate that air pollution significantly increases credit costs, supporting Hypothesis 1. According to environmental psychology theory, air pollution leads rating analysts to worry about their physical and mental health, thereby assigning pessimistic credit ratings to firms in polluted areas, which in turn raises corporate credit costs.

5.3. Mechanism Test of Credit Ratings

The theoretical analysis above indicates that air pollution increases corporate credit costs by lowering the credit ratings assigned by rating analysts to firms. Therefore, this paper employs a mediating effect model to test the credit rating mechanism. Specifically, using credit rating data from the Wind Database, this paper assigns numerical values to credit ratings through a quantitative method: AAA is assigned a value of 5, AA a value of 4, A a value of 3, BBB a value of 2, BB a value of 1, and CCC and below a value of 0, resulting in series data on corporate credit ratings (Creditscore). To test the mediating effect of credit ratings (Creditscore), the following regression models are successively introduced on the basis of Model (1):

$$Credit\mathrm{s}cor{e}_{i,t}=c+\alpha Airp+{\displaystyle \sum _j{\alpha }^j{X}_{i,t}}+{\gamma }_t+{\theta }_i+{\epsilon }_{i,t}$$

(2)

$$Credit{c}_{i,t}=c+\lambda Airp+\eta Creditscore+{\displaystyle \sum _j{\alpha }^j{X}_{i,t}}+{\gamma }_t+{\theta }_i+{\epsilon }_{i,t}$$

(3)

Model (2) is used to test the impact of air pollution on corporate credit ratings, and Model (3) is used to test the joint impact of credit ratings and air pollution on corporate credit costs. The results of the mechanism test are shown in

Table 5. Mechanism test of credit ratings.

Variables	(1)	(2)	(3)
	Creditc	Creditscore	Creditc
Airp	0.0134 **	−0.0064 *	0.0127 **
	(2.2600)	(−1.9537)	(2.1391)
Creditscore			−0.1150 ***
			(−3.2030)
Constant	3.9648 ***	−2.4689 ***	3.6810 ***
	(7.1976)	(−7.6900)	(6.6603)
Controls	Yes	Yes	Yes
Year FE	Yes	Yes	Yes
Indus FE	Yes	Yes	Yes
Obs	13,124	13,124	13,124
Adj R2	0.4237	0.9339	0.4252

Note: ***, ** and * indicate significance at the 1%, 5% and 10% levels, respectively. The t values in parentheses are all adjusted for clustering at the individual level.

. Column (1) presents the benchmark regression results, which were explained earlier and are not repeated here. Columns (2) and (3) report the results of the credit rating mechanism test. In Column (2), the regression coefficient of Airp is negative and significant at the 10% level, indicating that air pollution significantly reduces corporate credit ratings. As shown in Column (3), when both air pollution and credit ratings are included in the model, the regression coefficient of the Creditscore is negative and significant at the 1% level, suggesting that the reduction in corporate credit ratings caused by air pollution leads to an increase in corporate credit costs. Moreover, the regression coefficient of Airp is positive and significant at the 5% level. These test results indicate that air pollution increases corporate credit costs by lowering corporate credit ratings.

5.4. Robustness Tests

5.4.1. Placebo Test

Theoretically, the effect of air pollution on corporate credit costs might also be influenced by other factors. To exclude this possibility, this paper conducts a placebo test by randomly constructing air pollution indicators. Specifically, first, the data on the degree of air pollution are randomly shuffled; second, the shuffled air pollution data are used to replace the air pollution variable employed earlier, and the sample is fed into Model (1) for multiple regression testing; third, the process above is repeated 500 times. Figure 2 reports the results of the placebo test for air pollution and credit costs. The regression coefficients show a symmetrical distribution centered around 0, with the mean value of the estimated coefficients significantly differing from the coefficient value of 0.0183 in Column (4) of Table 3. In terms of the p value distribution, the p values of the regression coefficients for most random samples are greater than 0.1, indicating no significant correlation. The placebo test results suggest that the relationship between air pollution and corporate credit costs in the benchmark regression is indeed valid, ruling out the influence of other accidental factors to a certain extent.

5.4.2. Coefficient Stability Tests

The benchmark regression model may have unobservable omitted variables. To address this issue, this paper uses coefficient stability tests, following Oster [48], to measure the potential bias. Two parameters, R_max and δ, are introduced, representing the maximum R-squared achievable with omitted variables and the selection proportion. The consistent estimate β* of the key variable depends on R_max and δ. If β* remains stable after adding the observed controls, then the omitted variables’ bias is limited. According to Oster’s criteria, if β* at R_max = 1.3 and δ = 1 is within Airp’s 95% CI or if |δ| > 1 when R_max = 1.3 and β* = 0, omitted variables are not a serious issue. As

Table 6. Coefficient stability tests.

Testing Method	Judgment Criteria	Actual Calculation Results	Does It Pass
(1)	β* (Rmax,δ) ∈ [0.0093, 0.0272]	β* (Rmax,δ) = 0.0261	Yes
(2)	|δ| > 1	δ = 2.1226	Yes

shows, the model passes both tests, indicating that omitted variables do not materially affect the results.

5.4.3. Instrumental Variable Method

Air pollution may increase credit costs, but if enterprises adopt measures to increase their R&D investment in green environmental protection to reduce their credit costs, these measures could improve air quality. Therefore, there may be an endogeneity problem of bidirectional causality between the degree of air pollution and corporate credit costs. To address this problem, this paper employs the instrumental variable (IV) method to control for the potential interference of endogeneity issues in the research conclusions. Specifically, the proportion of raw coal in regional energy consumption (Coalpro) is selected as an instrumental variable for air pollution because, since 2016, some cities in China have begun to ban coal use [11]. The proportion of raw coal directly affects air pollution but has no direct link to corporate credit costs. Further tests on the instrumental variables show that the rk LM test result is significant at the 1% level and that the Wald F statistic is greater than the critical value of the Stock–Yogo test at the 10% significance level, indicating that both instrumental variables are correlated with the endogenous explanatory variable.

The results of the IV method test are shown in Columns (1) and (2) of

Table 7. Results of other robustness tests.

Variables	Instrumental Variable Method		Entropy Balancing	Replacing the Measurement of Corporate Credit Costs	Removing Interfering Samples			Two-Dimensional Clustering	Control for Omitted Variables
	First	Second			Removing Data from Municipalities	Removing Data During the COVID-19 Pandemic	Removing Data from Mining Industry Samples
	(1) Airp	(2) Creditc	(3) Creditc	(4) Creditc_D	(5) Creditc	(6) Creditc	(7) Creditc	(8) Creditc	(9) Creditc
Coalpro	29.4497 ***
	(20.9175)
Airp		0.0211 **	0.0173 ***	0.0241 ***	0.0173 ***	0.0196 ***	0.0208 ***	0.0183 ***	0.0129 ***
		(2.2733)	(3.3816)	(3.9691)	(3.7110)	(3.7469)	(4.3855)	(3.4947)	(2.8062)
Envpen									0.0880 ***
									(2.7843)
Ftrade									−0.0087 ***
									(−5.1210)
Egf									0.0223 ***
									(8.3723)
Constant	−3.0616	4.6173 ***	3.4181 ***	3.9554 ***	2.9239 ***	3.4112 ***	3.0588 ***	3.0829 ***	3.0427 ***
	(−1.6091)	(11.6099)	(7.4602)	(7.3359)	(6.6398)	(7.5384)	(7.5006)	(6.1787)	(7.7615)
Controls	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Indus FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Obs	33,524	33,524	33,524	33,524	26,471	27,244	32,689	33,524	33,524
Adj R2	0.3260	0.3260	0.3242	0.1770	0.3406	0.2792	0.3270	0.3261	0.3335

Note: *** and ** indicate significance at the 1%, and 5% levels, respectively. The t values in parentheses are all adjusted for clustering at the individual level.

. In the first-stage regression, the coefficient of the Coalpro term is positive and significant at the 1% level, indicating that the proportion of raw coal in a region significantly exacerbates air pollution. In the second-stage regression, the coefficient of the Airp term is positive and significant at the 5% level, suggesting that after controlling for endogeneity, air pollution still has a significant positive effect on corporate credit costs.

5.4.4. Entropy Balancing

Given that the propensity score matching (PSM) method heavily relies on the specification of the first-stage logit model and may lead to a certain loss of sample information, this paper employs the entropy balancing (EB) method to address potential biases arising from sample selection issues. Specifically, the EB method is used to eliminate imbalances in covariates between firms below and above the median air pollution level. As shown in Column (3) of Table 7, the coefficient of Airp remains positive and significant at the 1% level, indicating that the main results are robust to controlling for sample selection bias.

5.4.5. Alternative Measure of Corporate Credit Costs

To mitigate potential biases arising from measurement errors in key variables, this paper employs an alternative measure of corporate credit costs. Specifically, the ratio of borrowing costs to total liabilities (Creditc_D) is used as a substitute metric. The results of re-estimating Model (1) are presented in Column (4) of Table 7. The coefficient of Airp remains positive and significant at the 1% level, confirming the robustness of this study’s main findings.

5.4.6. Removing Interfering Samples

To exclude the interference of confounding samples on the relationship between air pollution and corporate credit costs, this paper successively conducts the following interference sample exclusion tests. First, samples from municipalities directly under the central government are excluded. These municipalities differ significantly from other regions in terms of their economic structure, policy support, and resource allocation. They typically have stricter environmental regulatory policies and more abundant environmental protection funding, while also possibly enjoying special treatment in credit resource allocation. This particularity may cause the relationship between air pollution and credit costs in the municipality samples to lack generality, thereby interfering with the research results of this paper. After excluding the municipality samples, the regression results of Model (1) are shown in Column (5) of Table 7, where the coefficient of the Airp term remains positive and significant at the 1% level. Second, the COVID-19 pandemic has had an unprecedented impact on the global economy and society. During this period, governments around the world introduced a series of unconventional economic stimulus and credit support policies to mitigate the negative impacts of the pandemic on enterprises. The implementation of these policies has altered corporate financing environments and credit cost structures while also affecting the production and operational activities of enterprises, thereby indirectly influencing air pollution levels. Therefore, the relationship between air pollution and credit costs in pandemic-period samples may be confounded by pandemic-related factors. The test results after excluding the COVID-19 pandemic-period samples are shown in Column (6) of Table 7, with no substantial changes in the results. Third, mining industry samples are excluded. Compared with other industries, the mining industry has unique sectoral characteristics, with significant differences in the manner and extent of environmental impacts from its production and operational activities. First, mining activities often directly damage the ecological environment, generating large amounts of waste and pollutants, making air pollution issues more prominent and complex; second, the mining industry is a capital-intensive sector, with special characteristics in terms of credit demand, financing channels, and the factors influencing credit costs. These sectoral characteristics may cause the relationship between air pollution and credit costs in the mining industry samples to deviate from the general patterns of other industries. After excluding the mining industry samples, the regression results of Model (1) are shown in Column (7) of Table 7, where the coefficient of the Airp term remains positive and significant at the 1% level.

5.4.7. Two-Dimensional Cluster Adjustment

To further alleviate potential issues of cross-sectional dependence in error terms across industries and time periods, this study re-estimated the regression using a two-way clustering approach that adjusts standard errors by both industry and year dimensions. The specific regression results are presented in Column (8) of Table 7. The coefficient of the Airp variable remains positive and significant at the 1% level, demonstrating that our core findings remain robust even after rigorously controlling for interindustry and intertemporal dependence through this dual-dimensional clustering adjustment. This evidence reinforces the validity of our research conclusions.

5.4.8. Control for Omitted Variables

Although our baseline model controls for firm-level and regional factors that may affect corporate credit costs, there may still be omitted variables that influence the results, potentially leading to estimation bias. To address this concern, we further include three additional control variables in the baseline model: environmental penalties (Envpen), foreign trade volume (Ftrade), and environmental governance expenditures (Egf). Specifically, Envpen is a dummy variable equal to 1 if the firm received an environmental penalty in that year and 0 otherwise; Ftrade is the natural logarithm of the firm’s export volume for the year; and Egf is the natural logarithm of the firm’s annual environmental governance expenditures. Column (9) of Table 7 reports the regression results with these additional controls. The coefficient on Airp remains significantly positive at the 1% level, indicating that our main findings continue to hold after addressing potential omitted variable bias.

6. Further Analysis

6.1. Economic Hypothesis Testing

The preceding theoretical analysis suggests that the premise underlying the impact of air pollution on lower corporate credit ratings is that credit rating analysts develop pessimistic emotions; however, this assumption has not yet been empirically tested. Therefore, this study employs rating downgrade probability and rating volatility as indirect proxies for analyst pessimism. The downgrade probability reflects analysts’ heightened risk aversion under the influence of negative emotions, leading to a greater tendency to lower credit ratings in order to avoid potential risk exposure or reputational consequences associated with overly optimistic assessments. This conservative behavior can be seen as a defensive judgment strategy in response to environmental uncertainty. In contrast, rating volatility captures the instability in credit assessments over time, which may arise from emotional fluctuations, cognitive fatigue, or impaired information processing. Exposure to adverse environmental stimuli—such as air pollution—can amplify these effects, making analysts more reactive to noise or ambiguous firm signals. This volatility is especially pronounced under conditions of information asymmetry or unclear firm risk profiles. In sum, downgrade probability reflects the directional pessimism in judgment, while rating volatility captures inconsistency in decision-making; together, they offer behavior-based measures of analyst sentiment that help illuminate the psychological channel through which air pollution may affect corporate credit costs. Specifically, this paper constructs Model (4) to examine the impact of air pollution on the pessimistic sentiment of credit rating analysts.

$$Em{o}_{i,t}=c+\psi Airp+{\displaystyle \sum _j{\psi }^j{X}_{i,t}}+{\gamma }_t+{\theta }_i+{\epsilon }_{i,t}$$

(4)

In Model (4), Emo represents the pessimistic sentiment of credit rating analysts and is measured in two ways. First, Emo1 refers to a rating downgrade: It takes the value of 1 if the credit rating assigned to the firm in the current year is lower than that of the previous year and 0 otherwise. Second, rating volatility is measured as the standard deviation of the firm’s credit ratings assigned by analysts over a three-year window (t − 1, t, t + 1). The results are reported in

Table 8. Economic hypothesis testing.

Variables	(1)	(2)
	Emo1	Emo2
Airp	0.0243 **	0.0014 *
	(1.9700)	(1.6602)
Constant	13.8487 ***	−0.0123
	(12.3598)	(−0.1967)
Controls	Yes	Yes
Year FE	Yes	Yes
Indus FE	Yes	Yes
Obs	13,124	13,124
Adj R2	0.2835	0.0198

Note: ***, ** and * indicate significance at the 1%, 5% and 10% levels, respectively. The t values in parentheses are all adjusted for clustering at the individual level.

. The regression coefficients of Aip are significantly positive at least at the 10% level, indicating that air pollution leads to credit rating downgrades and increases rating volatility. These findings suggest that air pollution induces pessimistic sentiment among credit rating analysts.

6.2. Moderating Effect Analysis

This paper examines the impact of air pollution on corporate credit costs and its underlying mechanisms, but the relationship between the two variables may also be influenced by industry characteristics and regional disparities. This study tests the moderating effects of high-polluting industries, manufacturing, and regional banking competition intensity on the relationship between air pollution and corporate credit costs. Specifically, an interaction term between air pollution (Airp) and the moderator variable (Moder) is further introduced into Model (1) to construct Model (5):

$$Credit{c}_{i,t}=c+\beta Airp+\varphi Airp\times Moder+\chi Mpder+{\displaystyle \sum _j{\alpha }^j{X}_{i,t}}+{\gamma }_t+{\theta }_i+{\epsilon }_{i,t}$$

(5)

Among them, the moderating variable Moder includes three variables: The first is high-polluting industries (Highpo). In accordance with the standards set by China’s Ministry of Environmental Protection in the “Categorized Management List for Environmental Verification of Listed Companies”, industries, such as B06, B07, B08, B09, C15, C17, C18, C19, C22, C25, C26, C27, C28, C29, C31, C32, D44, and D45, are classified as high-polluting industries. Highpo is assigned a value of 1 for these industries and 0 otherwise. The second variable is the manufacturing industry (Ismanuf). If an enterprise belongs to Industry C, as classified by the China Securities Regulatory Commission (CSRC) in 2012, Ismanuf is assigned a value of 1 and 0 otherwise. The third variable is the level of regional banking competition (HHI), which is measured by the Herfindahl–Hirschman Index of the banking industry. This index is a negative indicator: the higher its value, the lower the level of banking competition.

The results of the moderating effect analysis are presented in

Table 9. Moderating effect analysis results.

Variables	(1)	(2)	(3)
	Creditc	Creditc	Creditc
Airp	0.0070	0.0103	0.0389 ***
	(1.2409)	(1.2770)	(3.7169)
Airp × Highpo	0.0166 **
	(2.0099)
Highpo	0.4074 ***
	(7.6240)
Airp × Ismanuf		0.0198 **
		(2.1067)
Ismanuf		−0.1508 ***
		(−2.7318)
Airp × HHI			0.1724 **
			(2.5430)
HHI			−1.7327 ***
			(−3.9719)
Constant	3.1639 ***	2.6553 ***	2.5967 ***
	(8.2067)	(6.3447)	(6.3024)
Controls	Yes	Yes	Yes
Year FE	Yes	Yes	Yes
Indus FE	Yes	Yes	Yes
Obs	33,218	33,524	33,524
Adj R2	0.3465	0.2919	0.3275

Note: *** and ** indicate significance at the 1%, and 5% levels, respectively. The t values in parentheses are all adjusted for clustering at the individual level.

. In Column (1), the coefficient of the Airp × Highpo term is positive and significant at the 5% level, indicating that air pollution has a more pronounced positive effect on credit costs for high-polluting enterprises than for low-polluting enterprises, supporting Hypothesis 2. In Column (2), the coefficient of the Airp × Ismanuf term is positive and significant at the 5% level, revealing that the impact of air pollution on credit costs is amplified for manufacturing enterprises, supporting Hypothesis 3. In Column (3), the coefficient of the Airp × HHI term is positive and significant at the 5% level, suggesting that as regional banking competition intensity increases, the positive effect of air pollution on corporate credit costs is significantly weakened, supporting Hypothesis 4.

6.3. Expansion Test

6.3.1. Long Term Trend Test of the Driving Effect of Air Pollution on Corporate Credit Costs

Given that a region’s industrial development pattern and geographical environment are difficult to change significantly in a short period of time, the regional air pollution situation is also unlikely to improve in the short term. Under such circumstances, does the promoting effect of air pollution on corporate credit costs exhibit a long-term trend? To answer this question, this paper examines the impact of air pollution on corporate credit costs in future one-period, two-period, and three-period horizons. The regression results are presented in

Table 10. Long-term trend test of the driving effect of air pollution on corporate credit costs.

Variables	(1)	(2)	(3)
	Creditct+1	Creditct+2	Creditct+3
Airp	0.0162 ***	0.0164 ***	0.0169 ***
	(3.4459)	(3.3233)	(3.4084)
Constant	2.9941 ***	3.2444 ***	3.3870 ***
	(7.0574)	(7.3342)	(7.3844)
Controls	Yes	Yes	Yes
Year	Yes	Yes	Yes
INDUS	Yes	Yes	Yes
Obs	27,657	23,903	21,150
Adj R2	0.3487	0.3395	0.3420

Note: *** indicate significance at the 1% levels, respectively. The t values in parentheses are all adjusted for clustering at the individual level.

. The regression coefficients of Airp with Creditc_t+1, Creditc_t+2, and Creditc_t+3 are all positive and significant at the 1% level, and the coefficients show an increasing trend over time. This finding indicates that the promoting effect of air pollution on corporate credit costs has a long-term trend, and this effect becomes more pronounced over time.

6.3.2. Air Pollution, Credit Costs, and Corporate Commercial Credit Allocation

Faced with the dilemma of increased credit costs caused by air pollution, what coping strategies do enterprises adopt? Research shows that commercial credit refers to the common credit relationship formed by enterprises due to deferred payments or advance receipts in normal business activities and commodity transactions. Compared with bank credit, commercial credit financing is relatively flexible in terms of acquisition and does not require cumbersome collateral procedures [49]. Furthermore, enterprises can obtain funding support more conveniently based on their cooperative relationships and transaction conditions with suppliers and do not need to pay capital costs. So, do enterprises adjust their commercial credit allocation to cope with the increase in credit costs caused by air pollution?

From the perspective of commercial credit financing, an increase in credit costs increases the cost for enterprises to obtain funds from banks. To maintain normal production operations and capital turnover, enterprises need to seek alternative funding sources [50]. As a form of short-term financing, commercial credit financing can partially meet corporate funding needs. Therefore, enterprises may tend to utilize more commercial credit financing, alleviating their own capital pressure by borrowing suppliers’ funds. From the commercial credit supply perspective, in the face of higher credit costs, enterprises will also adopt a more cautious approach to providing commercial credit externally, reducing credit sales limits and terms for downstream customers [51]. This will be completed to ensure that they have sufficient funds to cover various expenses and debt repayments, reflecting enhanced risk prevention awareness. When confronted with increased credit costs and operational risks caused by air pollution, enterprises become more sensitive to risks. Providing commercial credit implies that enterprises must assume the credit risk of downstream customers. Under heightened operational uncertainties, enterprises will more cautiously assess the creditworthiness of downstream customers and reduce the trade credit supply to mitigate potential bad debt risk.

Therefore, this paper hypothesizes that enterprises may adopt a combined strategy of increasing commercial credit financing and reducing the commercial credit supply to cope with the challenge of high credit costs triggered by air pollution. To validate this hypothesis, taking the median of corporate credit costs as the classification criterion, this paper conducts group tests on the impact of air pollution on corporate commercial credit financing and the commercial credit supply to reveal the differences in this impact between the low credit cost group and the high credit cost group.

$$CCF{(CCS)}_{i,t}=c+\beta Airp+{\displaystyle \sum _j{\alpha }^j{X}_{i,t}}+{\gamma }_t+{\theta }_i+{\epsilon }_{i,t}$$

(6)

Among them, CCF (commercial credit financing) is measured by the ratio of accounts payable to operating income. A larger value indicates a higher scale of trade credit financing by enterprises; CCS (commercial credit supply) is measured by the proportion of the sum of accounts receivable, notes receivable, and prepayments to total assets. A larger value of this indicator represents a stronger capacity for commercial credit supply by enterprises. Columns (1) and (2) of

Table 11. Air pollution, credit costs, and corporate commercial credit allocation.

Variables	CCF		CCS
	(1) High Credit Cost Group	(2) Low Credit Cost Group	(3) High Credit Cost Group	(4) Low Credit Cost Group
Airp	0.0027 ***	0.0016	−0.0025 ***	−0.0016 **
	(3.4129)	(1.5616)	(−4.5322)	(−2.1661)
Constant	0.2744 ***	0.3045 ***	0.6459 ***	0.5283 ***
	(3.8877)	(4.0303)	(12.8457)	(9.7472)
Controls	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes
Indus FE	Yes	Yes	Yes	Yes
Obs	13,230	13,222	16,239	16,233
Adj R2	0.2991	0.3077	0.2475	0.2497
Chow Test	22.30 ***		18.68 ***

Note: *** and ** indicate significance at the 1%, and 5% levels, respectively. The t values in parentheses are all adjusted for clustering at the individual level.

report the results of the group tests on the impact of air pollution on corporate commercial credit financing. In Column (1), the coefficient of the Airp term in the high credit cost group is 0.0027 and is positive and significant at the 1% level. In Column (2), the coefficient of the Airp term in the low credit cost group is 0.0016 and is not significant. The test for differences in coefficients between groups is also significant at the 1% level. These results indicate that, compared with enterprises with lower credit costs, air pollution significantly promotes commercial credit financing for enterprises with higher credit costs. Columns (3) and (4) of Table 11 report the results of the group tests on the impact of air pollution on the corporate commercial credit supply. In Column (1), the coefficient of the Airp term in the high credit cost group is −0.0025 and is negative and significant at the 1% level. In Column (2), the coefficient of the Airp term in the low credit cost group is −0.0016 and is negative and significant at the 5% level. The test for differences in coefficients between groups is also significant at the 1% level. These results show that, compared with enterprises with lower credit costs, air pollution has a more pronounced inhibitory effect on the commercial credit supply for enterprises with higher credit costs. In summary, the findings indicate that enterprises tend to adopt a combined strategy of increasing commercial credit financing and reducing the commercial credit supply to cope with the dilemma of increased credit costs caused by air pollution.

7. Research Conclusions, Policy Recommendations, and Limitations

7.1. Research Conclusions

Currently, with global climate risks escalating and regional air pollution becoming increasingly severe, people’s physical and mental health is suffering significant harm. Meanwhile, whether enterprises face higher credit costs because rating analysts assign pessimistic credit ratings has become a critical issue, warranting attention. In this context, this paper takes Chinese A-share listed companies in the capital market from 2007 to 2022 as a sample to empirically examine the impact of air pollution on corporate credit costs.

The research results show that air pollution has a significant effect on increasing corporate credit costs, which is grounded in the fact that air pollution endangers the physical and mental health of rating analysts, leading them to assign more pessimistic credit ratings to enterprises located in areas with severe air pollution. The moderating effect analysis reveals that in high-polluting industries, manufacturing industries, and regions with lower banking competition, air pollution has a more pronounced role in increasing corporate credit costs. Further exploration indicates that when enterprises face the challenge of rising credit costs triggered by air pollution, they often adopt a combined strategy of increasing commercial credit financing and reducing the commercial credit supply. Although this strategy can alleviate corporations’ own funding shortages, it may have an adverse effect on supply chain stability.

7.2. Discussion

The findings of this study contribute to both the literature and practice by uncovering a novel channel through which environmental risks—specifically air pollution—affect corporate financing outcomes. While existing studies have emphasized the impact of air pollution on operational costs and labor productivity, this paper extends the literature by demonstrating that air pollution can indirectly increase firms’ credit costs through analyst behavior. In particular, our identification of rating downgrades and rating volatility as behavioral responses to environmental stress enriches the understanding of how non-financial externalities are priced in capital markets.

This study also provides empirical evidence that the impact of air pollution on credit costs is not uniform across contexts. Firms in high-polluting industries, manufacturing sectors, and regions with limited banking competition bear a disproportionately higher burden. This heterogeneity suggests that the financial system may unintentionally penalize firms exposed to environmental disadvantages, reinforcing capital constraints in areas already under stress.

Moreover, the observed corporate response—shifting toward commercial credit financing while reducing trade credit supply—highlights how financial pressure can lead firms to adopt strategies that may ease short-term liquidity needs but jeopardize long-term supply chain stability. This underscores the importance of understanding firm behavior under environmental-financial stress, particularly as climate and pollution-related risks become increasingly material in financial markets.

These findings point to the need for more nuanced environmental risk management within credit markets, including more transparent rating disclosures, better analyst working conditions, and targeted policy interventions.

7.3. Policy Recommendations

Based on the research findings above, this paper makes the following recommendations. First, the market should focus on credit rating analysts’ health protection and information disclosure. Given that air pollution affects the physical and mental health of rating analysts and thus influences corporate credit ratings, the market should attach great importance to analysts’ well-being. On the one hand, enterprises and institutions should equip analysts with high-quality anti-pollution equipment, such as professional face masks and air purifiers, and regularly organize health check-ups and psychological counseling. On the other hand, a more transparent rating information disclosure mechanism that details the impact weight and basis of external factors, such as air pollution, on the rating results in rating reports should be established. Doing so will enable market participants to more clearly understand the rating logic, reduce abnormal fluctuations in credit costs caused by information asymmetry, and stabilize the corporate financing environment.

Second, differentiated credit policies and industry support should be implemented. In response to the more significant impact of air pollution on high-polluting industries, manufacturing industries, and regions with low banking competition, regulatory authorities can guide financial institutions to implement differentiated credit policies. For high-polluting industries and manufacturing enterprises, under the premise of risk control, financial institutions should appropriately increase credit limits and reduce interest rates to encourage enterprises to upgrade their environmental protection technologies and reduce their pollution emissions. Meanwhile, in regions with low banking competition, governments can provide low-cost funds through policy banks or special support funds to sustain local enterprises. Additionally, efforts can be made to establish industry mutual aid funds to collectively address the pressure of credit costs caused by air pollution and reduce overall industry risks.

Third, enterprises should be guided to optimize their financing structures and supply chain management. Facing the challenge of rising credit costs triggered by air pollution, enterprises should actively optimize their financing structures. In addition to increasing commercial credit financing, they can explore diversified financing channels, such as issuing green bonds and introducing strategic investors, to reduce their reliance on single financing methods. At the same time, enterprises must prioritize supply chain stability by establishing long-term and stable cooperative relationships with upstream and downstream partners and strengthening information-sharing and risk-sharing mechanisms. When reducing the commercial credit supply, enterprises should communicate and negotiate with partners in advance, balancing their own capital needs and supply chain stability through measures such as extending payment cycles and offering preferential terms to achieve common sustainable development with supply chain partners.

7.4. Limitations

Despite this paper’s systematic examination of the impact of air pollution on corporate credit costs and the relevant mechanisms, some limitations remain. First, the long-term effects of corporate responses and their impact on the supply chain have not been thoroughly explored. This study shows that companies adopt a strategy of increasing commercial credit financing while reducing their supply to deal with the rise in credit costs caused by air pollution. This strategy may negatively impact supply chain stability. However, this paper does not research in depth the long-term effects of this strategy or the supply chain impact. It also fails to analyze whether corporate financing strategies will adjust in the long term under air pollution pressure and whether the supply chain will undergo structural changes, such as selecting new suppliers or reorganizing supply chain relationships. Future research could use data with a longer time span and apply case studies, dynamic simulations, etc. to track the evolution of corporate responses and assess long-term supply chain stability, offering more forward-looking references for businesses and policy-makers.

Second, although this paper studies the influence of high-polluting industries, manufacturing industries, and banking competition on the relationship between air pollution and corporate credit costs, it does not delve into the differences in the impact of air pollution on credit costs across enterprises of varying sizes and ownership types. There are significant differences among enterprises of different sizes in terms of their financial strength, risk-bearing capacity, and market influence. Large enterprises might more easily diversify their financing channels to mitigate the risk of rising credit costs due to air pollution, whereas small enterprises could face greater financing difficulties. Similarly, enterprises with different ownership types also vary in terms of financing policies and bank credit preferences. State-owned enterprises may have a government-backed credit advantage in obtaining loans, whereas private enterprises may be more vulnerable to external factors such as air pollution. Future research should further categorize enterprises by size and ownership and use group-based regression methods to compare the credit cost differences among various types of enterprises under the impact of air pollution, providing a basis for more targeted policies.

Third, this paper focuses on the effect of air pollution on corporate credit costs but ignores the potential interplay between other macroeconomic variables and air pollution. In the real economy, macroeconomic indicators, such as interest rates, inflation, and economic growth, can directly affect corporate credit costs and may have complex relationships with air pollution. For example, during an economic recession, air pollution might ease due to reduced production, but the credit market could tighten, increasing corporate credit costs. In this case, the impact of air pollution on corporate credit costs might be masked or amplified by macroeconomic factors. Future research should incorporate more macroeconomic variables into the analytical framework and use methods, such as structural equation modeling or simultaneous equations, to explore in depth the interactive mechanisms between air pollution and other macroeconomic variables to more accurately isolate the independent effect of air pollution on corporate credit costs.