Private Entrepreneurs of Communist Party Members and Corporate Social Responsibility: Evidence from Chinese Private Enterprise Survey

This paper examines whether private entrepreneurs with communist party member status affect corporate social responsibility (CSR) performance of private enterprises. Our empirical result reveals that firms led by private entrepreneurs of communist party members are significantly associated with higher CSR score. The result is robust to confronting several sources of endogeneity, e.g., re-examining the result using propensity score matching (PSM), entropy balancing (EB) and Heckman two-stage estimation, removing the influence of local religious culture. Further, the observed positive relation is particularly sharper for firms with party organization embedding and a good external legal environment. Our findings highlight entrepreneur political affiliation as an important driver of socially responsible corporate decision making.


Introduction
Corporate social responsibility (CSR) reflects the extent to which a firm actively engages in responsible business activities and has been considered to be important determinant of firm's sustainability. In China, the number of private enter-How to cite this paper: Li, J. Y., Liu, W. Q., & Yuan, D. L. (2020). Private Entrepreneurs of Communist Party Members and Corporate Social Responsibility: Evidence from Chinese Private Enterprise Survey. Theoretical Economics Letters prises accounts for more than 90% of the total number of enterprises and they are particularly important to the development of the national economy in terms of employment, tax contributions, and technological innovation 1 . As a result, the topic on the drivers of CSR for private enterprises is becoming interesting to scholars. Some scholars have identified factors that stimulate private enterprises to be involved in CSR activities, including political participant of entrepreneurs (Liang et al., 2010), business owners' social status (Gao et al., 2011), family involvement (Chen and Chen, 2014;Zhu, 2015), industrial associations (Chen et al., 2017) and party organization embedding (Xu et al., 2019). However, the role of the culture behind entrepreneurs' cognition and values is neglected.
China is an ideal laboratory for studying culture and finance. China not only has profound traditional cultures such as Confucianism, Taoism, and Buddhism, but also has communist culture (or red culture) with a rich revolutionary spirit and a heavy historical and cultural connotation. At the macro level, the impact of communist culture on China's macroeconomic development is obvious. The socialist country has become the second largest economy in the world. At the micro level, communist party members are an important carrier of the communist culture, which provides ideas for the study on corporate social responsibility and political culture with Chinese characteristics from a micro perspective.
In this paper, we will focus on the question of whether private entrepreneurs of communist party members carrying communist culture have an effect on CSR of private enterprises.
We conjecture that communist party member status is a driver of CSR in the private enterprises. China is a socialist country led by the Communist Party. The purpose of the Communist Party of China is to serve the people wholeheartedly, which has an internal connection with corporate social responsibility behavior that emphasizes responsibility to the society, environment, and employee. Therefore, private entrepreneurs who are members of the Communist Party will also implement the party's purpose. In addition, they are constrained to make more socially responsible decisions due to higher attention and higher moral requirement to communist party members from people. Consequently, we hypothesize that private entrepreneurs of communist party members are positively associated with CSR performance.
We use the data on private enterprise from Chinese Private Enterprise Survey (CPES) to confirm our hypothesis. We use a comprehensive CSR proxy including philanthropy, environmental protection and employee training to capture socially responsible activities of private enterprises. We identify the communist party member status of private entrepreneur based on whether he or she joins the Communist Party of China (CPC). We find a positive and significant association between private entrepreneurs of communist party members and CSR.
This result is robust to control for firm-level heterogeneity, entrepreneur-level 1 See the report on position and role of private enterprises in the national economy from the United Front Work Department of CPC Central Committee. http://www.zytzb.gov.cn/fgjjxw/108015.jhtml Theoretical Economics Letters heterogeneity, industry effect, region effect and year effect. Using propensity score matching (PSM), entropy balancing (EB), Heckman two-stage estimation specifications, and removing the influence of local religious culture, we find that our result and inference is robust to endogeneity concerns. Further, we conduct extended analyses from the internal party organization embedding and external legal environment. We find that the positive relation is stronger for private enterprises with party organization and a good external legal environment.
Our paper makes several contributions. First, it expands the research on the localization of upper echelons theory. We include the communist party member status with Chinese characteristic in the personal characteristics of entrepreneurs. Second, it builds on the growing literation that explores the impact the political culture on financial decision-making (one facet of which is the investment in socially responsible activities). Third, it is a useful supplement to the literature related to the informal system such as culture replacing the formal system. The paper proceeds as follows. Section 2 reviews previous researches on the characteristics of managers and CSR and develops our hypothesis. Section 3 describes the data and methodology and provides sample descriptives. Section 4 outlines the main results on the effect of private entrepreneurs of communist party members on CSR and discusses the potential endogeneity. Section 5 presents extended analyses on the effect. Section 6 concludes.

Literature Review and Hypothesis Development
Upper echelons theory creatively developed by Hambrick and Mason (1984) is the theoretical foundation for studying the personal characteristics of managers and corporate social responsibility (CSR). According to this theory, the demographic characteristics of managers will affect their cognitive basis and values, and then affect the decision-making behaviors and strategic choices of enterprises.
Based on upper echelons theory, the relevant study of corporate social responsibility has accumulated rich results. On the one hand, studies at the population physiological characteristics level generally link executives' gender or age with corporate social responsibility. For example, Wang & Coffey (1992) find that the proportion of female is positively and significantly associated with firms' charitable contributions, an important dimension of CSR. Manner (2010) finds that strong or exemplary corporate social performance, as measured by the strengths categories of KLD's ratings, is positively related to the CEO being female. Similarly, Xu and Li (2018) also find that firms with female CEO or higher ratio of female executives are more likely to improve the quality of CSR. These are in line with the views of Galbraith and Stephenson (1993) and Hillman and Dalziel (2003). In essence, both of them believe that female executives keep a watchful eye on ethics, and the human care of female executives is conducive to providing enterprises with more key resources and ability support related to ethics value. Aguinis and Glavas (2012) argue that younger generations are interested in CSR J. Y. Li et al. DOI: 10.4236/tel.2020.103041 638 Theoretical Economics Letters because of firm reputation. This is in stark contrast to the evidence in Zhang and Ji (2013) which indicates that the correlation between entrepreneurs' age and the social responsibility information level is significantly positive. However, Oh et al. (2016) do not find the main effect of CEO age on CSR.
On the other hand, studies focus more on the impact of executive's acquired experience on corporate social responsibility. Zhang and Ji (2013) find that the correlation between entrepreneurs' education and the social responsibility information level is significantly positive. Specifically, corporate social performance is better for the CEO having a bachelor's degree in humanities than the CEO having a bachelor's degree in economics (Manner, 2010). Muttakin et al. (2018) find that CEO power using a "power index" that comprises CEO duality, ownership, tenure and family CEO status is negatively associated with CSR disclosures. A recent study by Cronqvist and Yu (2017)  Reputation restriction may also urge private entrepreneurs of communist party members to undertake corporate social responsibility. The communist party members play an image of paragon and vanguard in the market activity. At the same time, they will also get more social attention and higher expectations from the masses (Gao et al., 2011). Under the dual restriction of personal reputation and corporate reputation, private entrepreneurs of communist party members are supposed to put more emphasis on their commitment and responsibility to stakeholders in the decision-making process, such as the decisions related to philanthropy donation, environmental pollution treatment, and employee training, so as to enhance the social responsibility of enterprises.
Based on the above discussion, private entrepreneurs of communist party members have higher ideological consciousness and stronger sense of identity with communist culture. Moreover, they will incorporate higher ethical standards into strategic decisions of companies under reputational constraints. Therefore, we conjecture that private entrepreneurs' communist party member status may be an important driver of socially responsible corporate decision making.
Accordingly, we propose the following hypothesis: Hypothesis: Private entrepreneurs with communist party member status are positively associated with CSR of private enterprises.

Data and Sample
We obtain data on private enterprise from Chinese Private Enterprise Survey For example, the second article of the Chapter I in Constitution of the Communist Party of China requests that members of the Communist Party of China must serve the people wholeheartedly, dedicate their whole lives to the realization of communism, and be ready to make any personal sacrifices. The third article requests that party member must to adhere to the principle that the interests of the Party and the people stand above everything else, subordinating their personal interests to the interests of the Party and the people, being the first to bear hardships and the last to enjoy comforts, working selflessly for the public interests and working to contribute more. Theoretical Economics Letters It is worth noting that CPES has been used in prior studies to discuss some issues on philanthropy (Liang et al., 2010;Gao et al., 2011;Chen and Chen, 2014), environmental innovation (Liao et al., 2019), or corporate social responsibility (Zhu, 2015;Xu et al., 2019;Chen et al., 2017) of private enterprises.
However, our paper is the first to use private enterprise survey data to empirically investigate the impact of private entrepreneurs of communist party members on corporate social responsibility. In addition, compared to previous studies that used only one-year data for analysis, we consider all years in which CSR data is available.

Variable Measurement
Dependent variable. For constructing proxy for corporate social responsibility (CSR), we rely on the following three questions in the questionnaire: 1) cash donated for public welfare undertakings such as poverty alleviation, disaster relief, environmental protection, charity, and cash converted by contribution in kind and engineering labour service; 2) the amount of investment in pollution abatement; 3) expenditure for staff training. We construct three dummy variables based on the answers to the above three questions. CSR_donation is a dummy variable equal to one if the answer to the first question is greater than zero and zero otherwise. CSR_environment is a dummy variable equal to one if the amount of investment in pollution abatement is more than zero and zero otherwise. CSR_employee is a dummy variable equal to one if the expenditure for staff training is more than zero and zero otherwise. The CSR used in the empirical analysis is the sum of these three dummy variables. Control variables. Following the executive background and CSR literature, we control for a rich set of entrepreneur characteristics and corporate characteristics that are likely to affect CSR of private enterprise. The usual control variables related to entrepreneur characteristics are gender, age, education, compensation, and political connection. Several standard corporate characteristics include firm's age, operation revenue, debt to assets ratio, profitability, and number of staff. In addition, we also control industry, region and year effect. A detailed definition of all the variables used in our analysis is provided in Table 1.

Regression Specifications
To assess how entrepreneurs of communist party members affect CSR of private enterprises, we base our estimations on both ordinary least squares (OLS) and ordered probit (OP). We use the following empirical specification: The dependent variable CSR measures the performance of private enterprises in philanthropy, environment protection, and staff training. Our main independent variable is ECPM, which measures private entrepreneur's communist party member status. The coefficient β captures the impact of private entrepreneurs of communist party members on corporate social responsibility. The remaining independent variables, included in Control, capture entrepreneur characteristics and corporate characteristics, as described in Table 1. , i r δ δ and y δ correspond to industry, region and year fixed effects, respectively. Sum of CSR_donation, CSR_environment and CSR_employee. The minimum value is zero and the maximum value is three for CSR. CSR_donation is a dummy variable equal to one if the amount of donation such as public welfare undertakings, contribution in kind and engineering labour service is greater than zero and zero otherwise. CSR_environment is a dummy variable equal to one if the amount of investment in pollution abatement is more than zero and zero otherwise. CSR_employee is a dummy variable equal to one if the expenditure for staff training is more than zero and zero otherwise.

Independent variable ECPM
Indicator that equals one if the entrepreneur is a communist party member and zero otherwise.

Control variables
Gender Indicator that equals one if the entrepreneur is a man and zero otherwise.

Age
Difference between the year filled in the questionnaire and the year of birth of the entrepreneur.

Education
The value of primary school education or below is zero. The value of junior high school or senior high school education is one. The value of college education or above is two.

Compensation
Natural logarithm of annual compensation of the entrepreneur (in units of ten thousand yuan).

PI
Political connection. Indicator that equals one if the entrepreneur is representative of National People's Congress or members of the CPPCC and zero otherwise.

Establish
Difference between the year filled in the questionnaire and the year of registration of private enterprise.

Revenue
Natural logarithm of operation revenue for private enterprise (in units of ten thousand yuan).

DTA
The percentage of total debt to total asset.

Profit
The percentage of operating profit to operation revenue.
Staff number Natural logarithm of the number of staff for private enterprise.

Industry
Twenty dummy variables were set according to the 20 industries involved in the questionnaire.

Region
Three dummy variables were set according to east, middle and west region. Year Calendar-year dummy.
This  This table reports descriptive statistics, correlation analysis, and difference test of mean and median. CSR is measured as total scores in three categories: CSR_donation, CSR_environment and CSR_employee. CSR_donation is a dummy variable equal to one if the amount of donation such as public welfare undertakings, contribution in kind and engineering labour service is greater than zero and zero otherwise. CSR_environment is a dummy variable equal to one if the amount of investment in pollution abatement is more than zero and zero otherwise. CSR_employee is a dummy variable equal to one if the expenditure for staff training is more than zero and zero otherwise. ECPM is an indicator that equals one if the entrepreneur is a communist party member and zero otherwise. Other variables are as defined in Table 1. In the correlation analysis, correlations in bold are statistically significant at p value ≤ 0.05. In difference test, ***, **, and * indicate significance at the 1%, 5%, and 10% level, respectively.

Baseline Result
We estimate the effect of private entrepreneurs of communist party members on CSR. The estimated results are presented in Table 3. Column (1) to (3) report OLS regression results and column (4) to (6) show the results of the OP. For comparison, we estimate three specifications: without the control variables, with the control variables on entrepreneur characteristics, with the control variables on corporate characteristics. In these specifications, we find that the coefficient on ECPM is positive and significant at 1% level, which confirms our hypothesis.
Both the OLS and OP results indicate that private entrepreneurs with communist party members status are significantly associated with higher CSR score.

Endogeneity Concern
Our results can be affected by endogeneity concerns resulting from potential selection bias and omitted correlated variables. To mitigate the endogeneity concerns and to check the robustness of our results, we use propensity score matching (PSM), entropy balancing (EB) and Heckman two-stage estimation to estimate Equations (1). We also exclude the impact of local religious culture on CSR of private enterprises.
We divide the sample into treatment group (ECPM = 1) and control group (ECPM = 0) to examine the causal effects of private entrepreneurs of communist party members on CSR for private enterprises.  This table reports both OLS and OP regression results of the effect of private entrepreneurs of communist party members on CSR, measured as total scores in three categories: CSR_donation, CSR_environment and CSR_employee. CSR_donation is a dummy variable equal to one if the amount of donation such as public welfare undertakings, contribution in kind and engineering labour service is greater than zero and zero otherwise. CSR_environment is a dummy variable equal to one if the amount of investment in pollution abatement is more than zero and zero otherwise. CSR_employee is a dummy variable equal to one if the expenditure for staff training is more than zero and zero otherwise. Independent variable EPCM is an indicator that equals one if the entrepreneur is a communist party member and zero otherwise. Other variables are as defined in Table 1. T-statistics are given in parentheses. ***, **, and * indicate significance at the 1%, 5%, and 10% level, respectively.    Table 1. T-statistics are given in parentheses. ***, **, and * indicate significance at the 1%, 5%, and 10% level, respectively.
Considering the possible sample selection bias between the treatment group and the control group, we further use the entropy balance (EB) method proposed by Hainmueller (2012) to perform weighted balance processing on the treatment and control group. Entropy balance can ensure that the treatment and control group achieve a balance on the sample characteristics, which will produce smaller coefficient bias and standard error than PSM. Table 5 reports the results from EM analysis. In column (1), we show estimates based on raw data, and do not control for any entrepreneur characteristics or corporate characteristics. The estimated coefficient on ECPM is 0.064 and statistically significant at the 5% level. In column (2) to (3), we add a set of standard entrepreneur and corporate controls one by one. And the significance level does not attenuate dramatically when we add controls for entrepreneur and corporate characteristics. The estimated coefficients on ECPM are 0.063 (t = 2.61) and 0.070 (t = 3.05), respectively. The evidence reveals a statistically significant and positive relation between entrepreneurs with communist party member status and CSR in Chinese private enterprises.
Another test for sample selection bias is Heckman two-step estimation, in which the first step adopts a probit model to predict the likelihood of private entrepreneurs of communist party members and calculates the predicted inverse Mills ratio (IMR), and the second step estimates the CSR performance using the IMR as a predictor in the model (1). We report the results of using Heckman two-step estimation in Table 6. Column (1) is the result of the first step regression and the dependent variable is ECPM. Column (2) to (4) are the results of the second step regression and the dependent variable is CSR. We add a set of standard entrepreneur controls in column (3) and corporate controls in column (4). Obviously, the coefficients on ECPM are positive and significant at the 1% level (0.071, 0.078 and 0.074). Significant and negative coefficients on IMR indicate the existence of sample selection bias. Again, this demonstrates that the communist party member status effect we focus on is statistically significant. In addition, we also exclude the impact of religious culture on CSR. It is important to emphasize that the communist party members have no religious belief. We only eliminate the influence of the religious culture in the region where the private enterprise is located, which is measured as the natural logarithm of the number of religious places in the region where the private enterprise is located. We report the results of excluding religious culture in Table 7. In column (1) to (5), we include the Buddhism, Taoism, Christian, Muslim, and Catholicism control, respectively. In column (6), we include all of the religion controls simultaneously. Clearly, the coefficients on the ECPM are positive and significant at the 1% level in all specifications.
In summary, the results in Tables 5-8 are consistent with the main results in Table 3, and hence our overall inference on the effect of private entrepreneurs of communist party members on CSR is robust to concerns about endogeneity.

Further Analyses
In this section, we test whether our results are more pronounced from the perspective of party organization embedding in the private enterprise and external legal environment. We measure the party organization embedding (Party) according to the presence or absence of party organization in private enterprise. We measure the legal environment (Legal) in the region where the private enterprise is located by using the legal environment index of China's provinces in Wang et al. (2018). We particularly focus on the coefficient on the interaction term between Party (or Legal) and ECPM. Similarly, we perform the OLS and OP regression as in Table 3 and the results are shown in Table 8. The column (1) and (3) present the OLS and OP regression results on the role of party organization embedding in the relation of private entrepreneurs of communist party members and CSR, respectively. Positive and significant coefficients on ECPM*Party (0.204 and 0.307) suggest that the effect of private entrepreneurs of communist party members on CSR is stronger when the party organization exists in the private enterprise. The column (2) and (4) present the OLS and OP regression results on the role of regional legal environment in the relation of private entrepreneurs of communist party members and CSR, respectively. We find significant and positive coefficients on the interaction term ECPM*Legal (0.030 and 0.042). This shows that the regional legal environment provides a good external environment for private entrepreneurs of communist party members to better fulfill CSR. That is, the effect of private entrepreneurs of communist party members on CSR is stronger with improvement of legal environment in the region where the private enterprise is located.

Conclusion
Based on the data from Chinese Private Enterprise Survey, this paper empirically investigates the impact of private entrepreneurs of communist party members on CSR. We find that private entrepreneurs with communist party member sta- This table reports the regression results of entropy balancing (EB) analysis. CSR is measured as total scores in three categories: CSR_donation, CSR_environment and CSR_employee. CSR_donation is a dummy variable equal to one if the amount of donation such as public welfare undertakings, contribution in kind and engineering labour service is greater than zero and zero otherwise. CSR_environment is a dummy variable equal to one if the amount of investment in pollution abatement is more than zero and zero otherwise. CSR_employee is a dummy variable equal to one if the expenditure for staff training is more than zero and zero otherwise. ECPM is an indicator that equals one if the entrepreneur is a communist party member and zero otherwise. Other variables are as defined in Table 1. T-statistics are given in parentheses. ***, **, and * indicate significance at the 1%, 5%, and 10% level, respectively.  (1) is the result of the first step regression and the dependent variable is ECPM, measured as an indicator that equals one if the entrepreneur is a communist party member and zero otherwise. Column (2) to (4) are the results of the second step regression and the dependent variable is CSR, measured as total scores in three categories: CSR_donation, CSR_environment and CSR_employee. CSR_donation is a dummy variable equal to one if the amount of donation such as public welfare undertakings, contribution in kind and engineering labour service is greater than zero and zero otherwise. CSR_environment is a dummy variable equal to one if the amount of investment in pollution abatement is more than zero and zero otherwise. CSR_employee is a dummy variable equal to one if the expenditure for staff training is more than zero and zero otherwise. IMR is the inverse Mills ratio, which is a control variable for the second step regression. Other variables are as defined in Table 1. T-statistics are given in parentheses. ***, **, and * indicate significance at the 1%, 5%, and 10% level, respectively. Theoretical Economics Letters This table reports the regression results of excluding religious culture. Column (1) to (5) control the Buddhism, Taoism, Christian, Muslim, and Catholicism, respectively. Column (6) includes all of the religion controls simultaneously. Buddhism and Taoism are measured as natural logarithm of the number of Buddhism temple and Taoist abbey in the region where the private enterprise is located. Christian, Muslim, and Catholicism are measured as natural logarithm of the number of Christian church, Muslim church, and Catholicism church in the region where the private enterprise is located. The dependent variable CSR is measured as total scores in three categories: CSR_donation, CSR_environment and CSR_employee. CSR_donation is a dummy variable equal to one if the amount of donation such as public welfare undertakings, contribution in kind and engineering labour service is greater than zero and zero otherwise. CSR_environment is a dummy variable equal to one if the amount of investment in pollution abatement is more than zero and zero otherwise. CSR_employee is a dummy variable equal to one if the expenditure for staff training is more than zero and zero otherwise. The independent variable ECPM is measured as an indicator that equals one if the entrepreneur is a communist party member and zero otherwise. Other variables are as defined in Table 1. T-statistics are given in parentheses. ***, **, and * indicate significance at the 1%, 5%, and 10% level, respectively. Theoretical Economics Letters