Do private firms invest more in environmental protection under political control? Evidence from China

Unlike the multi-party electoral systems in Western countries, China is characterized by unified Party and state structure. The CPC holds the dominant position in ruling roles. The Party and State are closely intertwined and difficult to separate from each other (Zhang et al 2022). The CPC appoints individuals to leadership positions in administrative bodies and governments, resulting in a closely connected 'party-government system' that governs the country's affairs.

The CPC has a hierarchical administrative structure consisting of central, regional, and local levels. This paper focuses on the CPC's grassroots organization, known as the local party committees, which are established at the firm level and serve as essential organs within a firm (Zhang et al 2022). Previously, local party committees only existed in state-owned enterprises. However, in the 1990s, in order to strengthen CPC leadership in the private sector, private firms were encouraged to establish local party committees (Shi et al 2014). By 2016, 1.86 million Chinese private firms had established local party committees, 10.4 times the number from a decade ago. These local party committees create networks through which the Chinese government can exert implicit influence on corporate behaviors.

A strand of literature has explored the various factors influencing corporate environmental behavior. These studies encompass a wide range of determinants, from media scrutiny (Li et al 2018) and the ramifications of external legal policies and regulations (Haque and Ntim 2018), to the pivotal role of entrepreneurs' cognition (Cai et al 2020, Gabler et al 2023), as well as distinctive corporate characteristics (Haller and Murphy 2012) and the implications of Employee Stock Ownership Plans (Kong et al 2024).

Specifically in the context of China, Boubaker et al (2024) highlighted that the country's anti-corruption campaign has fostered corporate environmental responsibility among energy firms. Similarly, Zhang et al (2023) identified that urban environmental legislation significantly bolsters corporate environmental operational performance, particularly in terminal governance, though it seemingly has negligible influence on corporate environmental management performance concerning process control. However, there is a scarcity of literature examining the role of local party committees in shaping corporate environmental behaviors. This paper seeks to bridge this gap by delving into this underexplored dimension. It enriches the existing literature on the determinants of corporate environmental behavior, thus offering a more comprehensive understanding of the intricate dynamics at play.

This study focuses on the local party committees within private enterprises, a distinctive structure within the Chinese context. Yet, its implications stretch far beyond, offering insights into comparable internal political structures in other emerging economies. Although not identical, these structures resonate with the role of local party committees. For example, trade unions in Mexico maintain close ties with the government or specific political parties, influencing corporate management, worker protection, and sometimes corporate decision-making (Lévesque and Murray 2005). Similarly, in India, certain trade associations are linked with political parties, swaying corporate policies and managerial decisions (Pandey and Varkkey 2020). Ratnam (2007) contends that the complementarity between Indian trade unions and the government has widespread impacts on public welfare. Thus, the role of local party committees in influencing corporate behavior warrants significant attention and should not be underestimated.

Studies from developed economies, which have distinct corporate governance structures and political affiliations compared to those in developing countries, indicate that political influence in democratic systems can have a beneficial impact on corporate social responsibility initiatives (Hasan and Jiang 2023). However, there is also critical examination of existing regulatory frameworks; for instance, Christensen (2022) challenges the effectiveness of relying solely on mandatory conflict mineral disclosures, currently the only relevant securities regulation in the U.S., and raises concerns about the adequacy of such reporting mandates in tackling broader issues like climate change.

Therefore, this paper's focus on local party committees emerges as a crucial point of interest. Firstly, it resonates with a wider implication, echoing the comparative analysis of similar structures in different global contexts as discussed earlier. Secondly, beyond the reliance on rigid top-down policies, the integration of these committees within firms introduces a more adaptable strategy. This approach can effectively bolster corporate compliance with regulations while circumventing the common pitfalls of policy resistance and resource misallocation that typically arise when firms employ countermeasures at the operational level.

Research on China primarily highlights the significant influence of the Communist Party of China (CPC) in shaping enterprise behavior (Chang and Wong 2004). For instance, the CPC membership has been identified as crucial in multiple dimensions of firm operations, international expansion (Marquis and Qiao 2020), bank loans access (Li et al 2008), agency conflicts (Li et al 2020), social responsibilities (Dong et al 2016). However, there is a deficiency in literature specifically addressing the integration of local party committees within firms. While a limited number of studies have touched upon the impact of local party committees, their focus has predominantly been on aspects like firm performance (Chang and Wong 2004) and business efficiency (Cao et al 2016), leaving a significant gap in empirical research concerning how these internal political structures shape corporate environmental practices. This study endeavors to fill this void by examining the influence of local party committees on environmental practices within private firms.

Existing studies have delivered inconclusive evidence on whether family firms have more environmental practices. On one hand, family owners, driven by the desire to preserve their socio-emotional wealth (SEW), may actively strive to build and maintain a reputable image, seeing environmental stewardship as a means to ensure the transgenerational transfer of their SEW and enhance their corporate image. Besides, family firms have a longer-term horizon for their business compared to non-family firms, which also suggests that reputation is of particular concern for family firms (Sah et al 2022), their commitment to environmental causes can serve as a powerful tool for building and maintaining a solid reputation. On the other hand, the 'dark side' of SEW highlights a potential conflict, where the prioritization of family interests may overshadow broader social responsibilities, thereby hindering socially responsible actions (Morck and Yeung 2004, Kellermanns et al 2012). This dichotomy portrays family firms as entities capable of exhibiting both socially responsible and irresponsible behaviors. Given this background, the role of local party committees in shaping the environmental practices of family-owned firms becomes particularly significant. This study contributes to the literature on the relation between family firms and environmental strategies by examining the interplay between ownership structure (family versus non-family firms) and the role of local party committees in shaping environmental strategies.

2.3.1. The local party committees

2.3.2. Environmental regulation

2.3.3. Family ownership

2.3.4. Party secretary

This study utilizes data from the Chinese Private Enterprise Survey, undertaken by the Privately Owned Enterprises Research Project Team in 2013. ⁴ This survey, one of a key national sampling survey in China, involves significant organizations like the United Front Work Department of CCP, All-China Federation of Industry and Commerce, State Administration for Market Regulation, Chinese Academy of Social Sciences, and the China Society of Private Economy. The data pool largely encompasses big and mid-sized firms, with a minor inclusion of individual household businesses, spanning across all 31 provinces of mainland China.

The CPES employed a multistage-stratified random sampling technique to ensure a comprehensive representation of every Chinese province and industry sector. The process involved several steps: initially determining the total sampling size and allocating quotas for each province, city, and autonomous region. This was followed by selecting a mix of cities and counties from each region, representing various economic development levels. The survey then balanced the number of urban and rural participants in line with the existing population distribution. Industry-specific samples were also carefully chosen to reflect the distinct economic sectors within these urban and rural areas. Finally, the actual firms for the survey were picked using a method that ensured equal representation across all intervals.

Data for this survey was meticulously gathered via comprehensive interviews within enterprises. Local Federations of Industry and Commerce, along with Bureaus of Industry and Commerce, deployed trained investigators to private enterprises. These investigators, who are well-versed in the workings of private enterprises, were responsible for distributing questionnaires, which were then completed by either the business owners or their key investors. In 2013, the survey targeted a sample representing $5.5{\rm{\unicode{x02031}}}$ of all private enterprises in China, issuing 7,000 questionnaires. Out of these, 6,144 were effectively returned, culminating in an impressive response rate of 87.8%.

The survey questionnaires consist of comprehensive personal profiles of business owners, including their family background, educational achievements, work history, and various information about their firms, such as firm size, age, financial data, and firm type. Of particular importance, the survey collects data on the total amount of money spent on environmental investments and pollution fees. It also includes information on whether or not a firm has established local party committees. This dataset is widely recognized as the most valuable resource for studying private firms in China and has been extensively used in previous research (Chen et al 2019, Li et al 2022, Long et al 2024).

Data on environmental regulation is collected manually from three statistical yearbooks: 'China Statistical Yearbook', 'China Environment Statistical Yearbook', and 'China City Statistical Yearbook'. The academic realm has not reached a consensus on a standardized metric for environmental regulation. Prominent methodologies in the field include the quantification of treatment levels for specific pollutants (Levinson 1996), and the ratio of total investment in pollution control to the total cost incurred by enterprises (Lanoie et al 2008). However, these indicators tend to focus on singular aspects of environmental regulation, introducing subjective biases and potentially skewing research outcomes. Consequently, this study, drawing inspiration from Cao et al (2020), adopts a more comprehensive index for the measurement of environmental regulation. In alignment with the data available, this paper incorporates three pivotal indicators: industrial wastewater discharge, industrial emissions of sulfur dioxide (SO₂), and industrial soot emissions. A higher environmental regulation index indicates higher pollution emissions and weaker environmental regulation. The calculation of the environmental regulation index is as follows:

First, we standardized the industrial wastewater discharge, industrial SO₂ emissions, and industrial soot emissions per unit of output value for each city:

$$\begin{eqnarray}&&U{E}_{{ij}}^{S}=\left[U{E}_{{ij}}-\min \left(U{E}_{j}\right)\right]/\left[\max \left(U{E}_{j}\right)-\min \left(U{E}_{j}\right)\right]\end{eqnarray} \tag{ 1 }$$

where $U{E}_{{ij}}^{S}$ is the standardized ${{UE}}_{{ij}};$ ${{UE}}_{{ij}}$ is the emissions per unit of output for city i and pollutants j; $\max \left(U{E}_{j}\right)$ is the maximum amount of emission for pollutants j among all cities. $\min \left(U{E}_{j}\right)$ is the minimum amount of emission for pollutants j among all cities.

The weights for different kinds of pollutants are:

$$\begin{eqnarray}{W}_{j}={UE}/\mathop{U{E}_{{ij}}}\limits^{̅}\end{eqnarray} \tag{ 2 }$$

where $\mathop{U{E}_{{ij}}}\limits^{̅}$ is the average level of emission per unit of output for pollutants j for all cities.

Finally, the environment regulation index is calculated as follows:

$$\begin{eqnarray}&&E{R}_{i}=\frac{1}{3}\displaystyle \sum _{j=1}^{3}{W}_{j}U{E}_{{ij}}^{S}\end{eqnarray} \tag{ 3 }$$

We drop the observations that do not have data for the key explanatory variable (Political_control). In China, we use the number of 8 employees as a criterion to differentiate between self-employed individuals and private enterprises. Therefore, we remove the firms that have less than eight employees. To minimize the impact of extreme observations, the continuous variables - firm age, sale, and profit - are truncated at 1%. Ultimately, the dataset consists of 4667 privately owned enterprises across 30 provinces in China.

The key explanatory variable being analyzed is political control, a binary variable that takes a value of 1 if the firm has local party committees and 0 otherwise. The dependent variable is the firm's investment in pollution abatement. Another dependent variable, PollutionControl, is a dummy variable that takes a value of 1 if the firm invests a positive amount of money in pollution reduction and 0 otherwise. This variable is used as an alternative dependent variable. EPI, or environment protection investment, is calculated as the firm's pollution abatement investment standardized by its sales. These measures of firms' pollution abatement investment represent proactive pollution prevention behaviors of firms, which are an essential part of their business activities and have a direct impact on improving pollution emissions and environmental management.

Drawing on existing literature research (Askildsen et al 2006, Hegde and Mishra 2019), our paper controls for a vector of firm-level and individual-level characteristics that could influence corporate environmental investment. The firm-level variables include firm size, firm age, and firm type. Firmsize is determined by the number of employees. ⁵ FirmAge is determined by the number of years since the firm establishment. The type of firm is determined by a categorical variable that classifies firms into sole proprietorship, partnership, limited liability company, and company limited by shares. In our analysis, we consider sole proprietorship as the reference group and exclude it from the regression.

At the individual-level, we include gender (Female), represented by a dummy variable that takes the value 1 if the entrepreneur is female and 0 otherwise. OwnerAge is determined by the age of the entrepreneur at the time of the survey. Education is measured by the number of years of schooling. Owner_PartySecretary is a dummy variable that takes the value 1 if the entrepreneur is a Party secretary and 0 otherwise. Additionally, we also control provincial and industrial fixed effects. For more detailed information on the variables and their descriptive statistics, please refer to table 1.

Despite the inclusion of numerous control variables in the regression, there may still be unobservable variables that could introduce bias. The existence of omitted variables, such as corporate culture, could potentially bias the causal impact of local party committees on corporate environmental investment. For instance, firms with a strong social responsibility culture may be more receptive to the integration of local party committees. Simultaneously, such a corporate culture would also naturally engage in higher environmental investment. Thus, not controlling for corporate culture may lead to an overestimation of the causal effect of local party committees. To mitigate the risk of omitted variable bias, this study further employs two alternative estimation methods: Propensity Score Matching and Instrumental Variable approaches. By comparing estimates from these two methods with our Ordinary Least Squares (OLS) results, we aim to reinforce the robustness of our conclusions, ensuring they are not unduly influenced by unobserved variables.

Table 1 shows that around 45% of firms have local party committees. The average age of private firms is approximately 10 years. Around 73% of private firms are classified as limited liability companies. The percentage of female private firm owners is only 14%. On average, firm owners have completed around 14 years of schooling and have an average age of 46. Furthermore, the data indicates that private firms display significant variations in terms of pollution abatement expenditure and EPI.

Table 2 displays the balancing checks performed to compare the treatment and control groups based on different variables at the owner-level and firm-level. Across all dimensions, there are notable disparities between firms with and without local party committees. On average, firms with local party committees tend to be younger, have fewer employees, and allocate less funds towards research and development. Furthermore, firms with local party committees exhibit higher proportions of female and younger entrepreneurs.

We want to examine whether political control has an effect on corporate environmental investment in general. The dependent variable is the natural logarithm of the firm's pollution abatement investment. Following Li et al (2008), we apply ordinary least-squares regression for analysis, with the equation as follows:

$$\begin{eqnarray}&&\begin{array}{c}{{pollution}{\rm{\_}}{control}{\rm{\_}}{invest}}_{{ijp}}={\beta }_{0}+{\beta }_{1}{{party}{\rm{\_}}{organization}}_{{ijp}}\\ +\delta {X}_{{ijk}}+{\beta }_{j}+{\beta }_{p}+{\varepsilon }_{{ijk}}\end{array}\end{eqnarray} \tag{ 4 }$$

Where, ${{pollution\_control\_invest}}_{{ijp}}$ is measured as the natural logarithm of 1 plus pollution abatement investment for firm i, sector $j,$ province $p$ in 2013. The ${{party\_organization}}_{{ijp}}$ variable is a dummy variable taking value 1 if local party committee is embedded in firm $i$ in sector $j$ and province $p,$ and 0 otherwise. The vector ${X}_{{ijkt}}$ is a set of firm's attributes and entrepreneur's attributes variables. We also control the vector of a province dummy ${\beta }_{p}$ and industry dummy ${\beta }_{j}.$ ${\varepsilon }_{{ijkt}}$ denotes the random noise term. We cluster the standard errors at the industry level to account for the correlation of the error terms across firms in the same industry. ${\beta }_{1}$ are our coefficients of interest, as they capture the relationship between local party committees and the firm's pollution abatement investment.

The results from the baseline estimation are presented in column 1 of table 3. These results indicate that there is a positive and statistically significant relationship between political control and the firm's expenditure on pollution reduction. This suggests that the presence of local party committees in the firm may have increased the firm's motivation to address pollution issues and invest more in pollution reduction.

In columns 2 and 3, we include different control variables as discussed in the previous section. In column 2, we include control variables related to owner-level characteristics. In column 3, we additionally include control variables related to firm-level characteristics. Despite a decrease in the estimated magnitude, we consistently find a positive and statistically significant effect of political control. These results support the testing of Hypothesis 1.

If other organizations are integrated within the company, the estimates of political control might mistakenly capture the effects of those confounding organizations instead of the impact of political control. Another notable organization in Chinese firms is labor unions. As noted by Fang and Ge (2012), these unions often act as an extension of the management and can sway company decisions, including those related to pollution control. In our data, around 38.5% of private firms had both local party committees and a labor union. About 34% had neither of these, while the rest, 27.5%, had one or the other.

When a firm houses multiple organizations, it's important to separate their individual impacts. If unions positively affect a firm's environmental investment and also coexist with political control, the observed effect might wrongly be attributed solely to political control. To tackle this, we introduced a control variable indicating the presence of a union in the firm. This adjustment allows for a more accurate comparison between firms with party committees and those without, under similar union statuses. In table 3, column 4, we observe that even with this control, political control maintains a significant and positive influence on a firm's spending on pollution control, though the impact is slightly reduced. This finding indicates that the effect of political control on environmental investment is genuine and not merely a reflection of the presence of a union in the firm.

To mitigate concerns about the endogeneity of political control and strengthen the validity of the results, a series of robustness checks are performed.

4.3.1. Change sample

4.3.2. Alternative measures of pollution abatement expenditure

4.3.3. Propensity score matching (PSM)

Our study might face a selection bias issue. Firms having local party committees may exhibit distinct characteristics compared to those without such committees. If these differences influence both the corporate environmental investments and political control, our study will suffer a self-selection problem. In order to alleviate the endogenous problems caused by this selection bias, we adopted the PSM method. This method has been widely used in academic research for reducing selection bias (see Ho et al 2023, Kong et al 2024).

The PSM method proceeds as follows. Initially, the full sample is divided into two groups: the treatment group, which consists of firms with local party committees, and the control group, which consists of firms without local party committees. Next, covariant variables that could potentially influence the likelihood of firms having local party committees are used to estimate the propensity score through a logit regression. The propensity score represents the probability of firms being in the treatment group. Based on the propensity score, the treatment group and the control group are matched using a 1:1 nearest neighbor matching with replacement. This matching process helps select a control sample that is closely matched to the treatment group.

The graph in figure 1 displays the kernel density function curves, comparing the propensity scores before and after matching. Observing figure 1(a), it is evident that the probability density distributions of the propensity scores for the two groups of samples differ significantly. The control group is inclined towards being left-skewed and more concentrated. However, after implementing nearest-neighbor matching, the probability density distributions of the two groups closely align (as shown in figure 1(b)). These results suggest that the characteristics of the two groups of firms become more similar after matching, indicating a substantial reduction in sample selectivity bias. Furthermore, the covariate test results presented in table 4 demonstrate that there were no significant differences in the mean values of the covariates between the treatment and control groups after performing propensity score matching. This indicates that the distribution of the variables became balanced between the two groups, validating the use of the PSM method.

Zoom In Zoom Out Reset image size

Table 4. The results of covariate balance checks.

	Treatment	Control	P-value
Female	0.100	0.092	0.453
Age	47.736	48.071	0.240
Eduy	15.015	14.932	0.363
Fage	11.351	11.596	0.226
Employment	5.135	5.132	0.948
Partner	0.035	0.031	0.606
Limited	0.757	0.753	0.797
Company	0.113	0.111	0.861
R&D	2.202	2.128	0.479

This table reports the results of covariate balance checks on the mean difference in the covariates used in the Logit model between the treatment firms and the control firms, matched on the PSM approach.

Table 5 presents the findings from the PSM analysis. In column 1, the results from the logistic regression underscore the significant role of key covariates in determining the presence of local party committees in firms. This indicates that the selected covariates effectively explain the likelihood of firms having local party committees. Column 3 presents the outcomes using matched samples. Here, it's evident that political control exerts a stronger influence on corporate environmental expenditure than initially observed in the baseline results from the full sample, which are detailed in column 2.

Table 5. The PSM robustness check.

Dependent variable	Political_control	Log(1+Pollution abatement Expenditure)
	(1)	(2)	(3)
	Logit	Full sample	Matched sample
Political_control		0.313***	0.419***
		(3.30)	(5.32)
Female	−0.183	−0.113**	−0.273***
	(−1.56)	(−2.20)	(−3.59)
Age	0.020***	−0.008***	−0.013**
	(4.09)	(−3.01)	(−2.84)
Eduy	0.099***	0.004	−0.024
	(5.18)	(0.47)	(−1.14)
Fage	0.044***	−0.001	−0.002
	(8.58)	(−0.10)	(−0.17)
Employment	0.552***	0.412***	0.414***
	(15.37)	(5.27)	(4.96)
Partner	0.250	0.074	−0.105
	(1.46)	(0.55)	(−0.63)
Limited	0.287***	−0.032	0.103
	(4.62)	(−0.75)	(0.87)
Company	0.427***	0.135	0.345*
	(4.74)	(1.66)	(1.80)
R&D	0.036***	0.137***	0.138***
	(3.29)	(12.77)	(12.43)
Industry fixed effect	Yes	Yes	Yes
Province fixed effect	Yes	Yes	Yes
Obs.	3499	3304	2942

*** denotes significance at 1%, ** at 5%, and * at 10%. All observations are at the firm level. Column 1 shows the logit regression to calculate propensity scores, and the dependent variable is political control. In columns 2 and 3, the dependent variable is the natural log of 1 plus expenditures on pollution reduction. Column 2 shows the full sample OLS regression. Column 3 shows regression results using matched sample. PSM method uses nearest neighbour matching within a calliper of 0.03 with replacement. All regression control for industry fixed effects and province fixed effects. The standard errors are reported in parentheses, clustered by industry.

4.3.4. Instrument variable approach (IV)

To further address the issue of endogeneity of political control, we conducted an instrumental variable analysis. The IV method is extensively employed in the literature to mitigate endogeneity concerns (Long and Yang 2016, Lei and Nugent 2018). Following Long and Yang (2016), we use political control intensity at the industry-province level as the instrumental variable. This refers to the proportion of firms within each province-industry cell. The validity of this instrument is represented as follows. Firstly, in provinces where political control intensity is higher, firms may be more likely to establish local party committees due to the overall environment. Secondly, the measure obtained from location-industry averages is influenced solely by the inherent characteristics of regions and industries, making it likely exogenous to a firm's expenditure on pollution reduction.

In the first step of the IV approach:

$$\begin{eqnarray}&&{{Party}{\rm{\_}}{organization}}_{{ijp}}={\beta }_{1}{{iv}}_{{jp}}+{\gamma }_{1}{X}_{{ijp}}+{\theta }_{p}+{\tau }_{j}+{\nu }_{{ijp}}\end{eqnarray} \tag{ 5 }$$

In the second step:

$$\begin{eqnarray}&&\begin{array}{c}Pollution\_Control\_Expenditur{e}_{ijp}=\alpha +\beta psrty\_{\widehat{organizationi}}_{ijp}\\ +\gamma {X}_{ijp}+{\theta }_{p}+{\tau }_{j}+{\varepsilon }_{ijp}\end{array}\end{eqnarray} \tag{ 6 }$$

The IV estimates are presented in table 6. In the first column, the first-stage results are reported. The coefficient of the instrument variable is significantly positive, which aligns with previous claims about the instrument's relevance in explaining whether a firm has local party committees. The second stage estimates are summarized in column 2. The coefficient of the instrument variable is positive and statistically significant, providing evidence for the significance of political control in promoting the company's investment in environmental pollution reduction. The magnitude of the instrument variable is larger than the estimate obtained through OLS. The F-test result of 254 indicates that there is no weak instrument problem. The Hausman test results reject the null hypothesis that the explanatory variables are exogenous. In order to ensure the robustness of the results against heteroscedasticity, GMM estimation results are included in the third column. The GMM estimates closely align with the instrument variable estimates. Overall, these consistent findings demonstrate that political control enhances a firm's expenditure on pollution reduction.

Table 6. The results of using instrument variables to reduce endogeneity issues.

Dependent variable	Political_control	Log(1+Pollution abatement Expenditure)
	(1)	(2)	(3)
	First stage	Second stage	GMM
IV	0.754***
	(0.044)
Political_control		1.565***	1.450***
		(0.404)	(0.199)
Province fixed effect	Yes	Yes	Yes
Owner-level control	Yes	Yes	Yes
Firm-level control	Yes	Yes	Yes
F-test	254
Hausman test (p-value)	0.00
Obs.	3499	3304	3580
R-squared	0.29	0.22	0.23

*** denotes significance at 1%, ** at 5%, and * at 10%. All observations are at the firm level. IV represents the instrument variable: the industry-province level local party committees' intensity. Column 1 and 2 shows the first and second stages of 2SLS regression. The reported F-test is for the null hypothesis that all coefficients are jointly equal to zero. The Hausman test is for the null hypothesis that all explanatory variables are exogenous. The alternative GMM regression is included in column 3. Firm-level control includes firm age, number of employments, firm type, and R&D investment. All regression control for industry fixed effects and province fixed effects. The standard errors are reported in parentheses, clustered by industry.

The pollution fee in China, levied by the environmental protection department, serves as a compulsory measure to combat pollution. This fee, unlike voluntary anti-pollution measures, is a mandatory cost for businesses. Despite this, non-compliance is notable. In 2015, the Ministry of Environmental Protection reported that 662 state-monitored enterprises across 29 provinces failed to pay sewage charges, resulting in unpaid dues of 696 million yuan. This failure undermines the government's intent to use the pollution fee as an incentive for reducing pollution. Some studies have used state ownership to indicate a firm's political affiliation and found that firms with state ownership pay lower environmental taxes (Maung et al 2016). Given the role of local party committees in overseeing and guiding firms, we aim to explore if the presence of these committees influences firms' adherence to environmental tax laws.

The results detailed in table 7 are consistently significant and positive across all three columns, underscoring the influential role of local party committees in motivating firms to fulfill their pollution fee obligations. This consistent pattern bolsters the idea that local party committees act as supervisors, aiding firms in paying pollution fees and subsequently increasing investment in pollution control.

This section explores where and how political control can exert a more significant impact. Table 8 presents the results of three moderators to the effect of political control. Column 1 shows how external environment regulation could moderate the effect of political control. Column 2 represents the family firm mediating effect. Column 3 reports the moderating effect of entrepreneurs being party secretaries.

Column 1 of table 8 reveals that while environmental regulation positively impacts a firm's spending on pollution control, this effect isn't statistically significant. Interestingly, the interaction term between environmental regulation and political control shows a significantly negative coefficient. This implies that political control tends to have a more substantial effect in regions with weaker environmental regulation. Essentially, the results from column 1 suggest that political control within private firms and external environmental regulation function as two distinct governance mechanisms, with a notable substitution effect between them. This finding provides empirical support for Hypothesis 2, indicating that in regions with weaker environmental rules, the role of political control in guiding firms towards environmental responsibility becomes more crucial.

In Column 2 of table 8, the influence of family ownership on a firm's pollution reduction investments, although positive, isn't statistically significant. This observation resonates with the tenets of the SEW and stewardship theories, which posit that family-owned firms tend to embrace pro-environmental behaviors, mirroring the findings of Agostino and Ruberto (2021). However, the interaction between family ownership and political control shows a significantly negative coefficient, suggesting that a stronger presence of family ownership might dilute the effectiveness of political control. This could be due to the intricate dynamics and potential conflicts within family-dominated firms, which make it challenging for political directives to penetrate and influence decision-making. Thus, Hypothesis 3 is supported.

Column 3 of table 8 presents an interesting dynamic: when an entrepreneur also holds the position of the Party secretary, the potency of political control in driving pollution reduction efforts is amplified. This underscores the synergy between entrepreneurial leadership and political commitment in steering a firm towards environmentally responsible practices, thereby lending credence to Hypothesis 4.