Abstract
CHINA JOURNAL OF ACCOUNTING STUDIES 2019, VOL. 7, NO. 2, 270–291 https://doi.org/10.1080/21697213.2019.1676050 The CPC party organization in privately controlled listed companies and earnings management a b a Dengjin Zheng , Deren Xie and Wei Yuan a b School of Accountancy, Central University of Finance and Economics, Beijing, China; School of Economics and Management, Tsinghua University, Beijing, China ABSTRACT KEYWORDS Communist party of China; This paper studies the real effect of the party organization of party organisation; privately- Communist Party of China (hereafter, CPCPO) on the privately con- controlled enterprises; trolled listed companies’ governance from the perspective of earn- earnings management ings management. We find a significantly negative relationship between the party organisation activities and earnings manage- ment. The possible mechanisms are to encourage controllers of those firms to participate in the activities of CPCPO and even serve as the secretaries, and to make directors whom are CPC party member enter audit committee. Furthermore, this relation- ship is more pronounced when the chairman or CEO is a CPC member, has political connections, and when the time is politically sensitive period. The relationship is also more pronounced in com- panies with higher agency costs, weaker information environments, and greater financing constraints. Different from previous literature focusing on the role of CPCPO in state-owned enterprises, this paper explores the governance role of CPCPO in privately con- trolled firms. 1. Introduction The embedding of party organisation of the Communist Party of China into privately controlled enterprises (hereafter, Non-SOEs) has become a common political and eco- nomic phenomenon in China. However, it is still in doubt whether CPCPO plays an effective role in privately controlled enterprises’ corporate governance. The Party Constitution of CPC (hereafter, Party Constitution) and The Company Law of China (here- after, Company Law) clearly stipulates that CPCPO is a formal institutional arrangement in Non-SOEs. The report of the 18th National Congress of the CPC specified that CPCPO in privately controlled enterprises must fulfil their duties of “carrying out CPC’s line and policy, supervising enterprises to comply with the laws and regulations, uniting the workers, and safeguarding the interests of all parties.’ In order to furtherly guide the party construction of Non-SOEs, the General Office of the CPC Central Committee issued CONTACT Dengjin Zheng zhengdj.13@sem.tsinghua.edu.cn School of Accountancy, Central University of Finance and Economics, Beijing, China Paper accepted by Kangtao Ye. A company should establish CPCPO and provide necessary conditions for CPCPO activities when the company has three or more communist party members. © 2019 Accounting Society of China CHINA JOURNAL OF ACCOUNTING STUDIES 271 the Opinions on Strengthening and Improving the Construction Work of Private Enterprises (hereafter, Opinions) in May 2012. The Opinion clearly defines the functional and orienta- tion of CPCPO in Non-SOEs. According to the 2017 Statistical Bulletin of the CPC issued by the Organisation Department of the CPC Central Committee, as of 31 December 2017, 1.877 million (73.1%) Non-SOEs have established CPCPOs, and the construction rate of CPCPO in those Non-SOEs which should establish CPCPO has reached 99% As a result, an important and open question is: Has CPCPO really played a significant governance role in Non-SOEs? At a press conference of the State Council Information Office in February 2017, the former chairman of China Securities Regulatory Commission (hereafter, CSRC) stated that ‘the CSRC and the China Association of Listed Companies are working with the relevant agencies to revise the corporate governance guidelines for listed companies and ask firms to clarify the structure, tasks and rules of CPCPO. In the Non-SOEs, CPC’s guidelines and the right of its members should be specified in accordance with the requirements of the Company Law, and both CPCPO and its members should play an effective role.’ Has CPCPO in Non-SOEs played a substantial role? Can it help to restrain the earning management and fulfil its responsibility to safeguard the interests of all parties? In theory, on the one hand, CPCPO’s discipline and the responsibility to safeguard the interests of all parties will subtly affect the behaviour of the controlling shareholders and senior executives and improve their self-discipline. Thus, CPCPO will restrain earnings management of Non- SOEs. On the other hand, the Non-SOEs might establish CPCPO just in form to meet the basic requirements by regulators. CPCPO may not have a substantial impact but stay in ‘formalism’. Therefore, an empirical question worth studying is: Does CPCPO help to restrain the earnings management of Non-SOEs? We manually collected the data of CPCPO in Non-SOEs listed companies to examine the restraining effect of CPCPO on earnings management. The results show that the power of CPCPO is significantly negative with earnings management. The one possible mechanism is to encourage the chairmen and CEOs of Non-SOEs to participate in the activities organised by CPCPO and even serve as the party secretaries, so as to enhance their understanding of the aims and spirit of CPC and thus help to restrain the earnings management. The other possible mechanism is to make directors whom are CPC party member enter the audit committee that is helpful to supervise a firm’s financial statement disclosure and internal control process, and improve accounting information quality. Furthermore, this relationship is more pronounced when the chairmen or CEOs are communist party members, have political connections, and when the time were politically sensitive periods. The relation is also more pronounced in companies with higher agency costs, weaker information environments and greater financing constraints. Finally, the conclusions of this paper is robust by using instrumental variables, DID model, PSM method and excluding several possible alternative explanations. Our study contributes to the literature in the following ways. Firstly, this paper enriches the literature of Non-SOEs’ CPCPO and the literature on the influence of political party on the economic consequences of enterprises. Prior literature mainly studies the economic consequences of CPCPO in state-owned enterprises’ (SOEs) (Chen & Lu, 2014; Ma, Wang, & Almost all of the Non-SOE listed companies in China have established CPCPOs. However, CPCPO in different Non-SOE listed companies may have different effects, which is also the focus issue of this paper.. 272 D. ZHENG ET AL. Shen, 2012, 2013). However, due to the different positioning and structure of CPCPO among SOEs and non-SOEs, the research conclusion based on SOEs may not be applicable to non-SOEs. Only a few papers study the consequence of CPCPO in Non-SOEs, but they focus on the impact of CPCPO on corporate donation behaviour and employee welfare, and only use an indicator variable from the database called ‘National Private Enterprise Sampling Survey Data’ to proxy the power of CPCPO (Liang, Chen, & Gai, 2010; Long & Yang, 2014). With the construction rate of CPCPO in the Non-SOEs being about 100%, this dummy variable is difficult to capture the different influence of CPCPO in Non-SOEs. This paper tries to measure the influence of CPCPO by using the number of activities organised by CPCPO and studies the impact of CPCPO on earnings management. Secondly, this paper complements the relevant literature on corporate governance with Chinese characteristics. The embedding of CPCPO in the Non-SOEs is a new and special corporate governance system under China’s institutional background. Whether CPCPO in Non-SOEs could have positive governance effect is an open question with great tension: CPCPO may promote the corporate governance to fulfil its aims, i.e. safeguarding the interests of all parties; it may also be used as a nominal ‘decoration’ or even a tool of ‘rent-seeking’. However, the existing literature discusses little about whether CPCPO affects the governance of Non-SOEs. And this paper extends the literature by exploring whether CPCPO plays a role in Non-SOEs’ governance from a perspective of earnings management. Thirdly, this paper also has some strong practical implications. The empirical evidences of this paper show that CPCPO in Non-SOEs does help to restrain the earnings manage- ment behaviour, which reflects the positive effect of the construction of CPCPO. Our evidence indicates that CPCPO plays a role in ‘implementing the CPC’s policy, guiding and supervising enterprises to abide by the laws and regulations’. Therefore, it has certain reference value for the CPC Central Committee to promote the construction of CPCPO in Non-SOEs. The remainder of the paper is organised as follows. Section 2 summarises the related literature and develops our hypotheses. Section 3 discusses our empirical design, sample and data source. Section 4 presents the results for our main tests, cross-sectional variation, and robustness tests. We conclude in Section 5. 2. Literature review and hypothesis development 2.1. Literature review In the past, the literature on government-corporate relations mostly focuses on the economic consequences of political connection or government intervention on enter- prises (Chen, Chen, & Dong, 2017;Fisman, 2001). Evidences in other countries on party- corporate relations are mainly based on party elections and study how political uncertainty affects corporate investment, corporate financing and other corporate behaviours (Goldman, Rocholl, & So, 2009;Julio & Yook, 2012). Domestic scholars have paid lots of attention to the influence of government on enterprise behaviour (Chen et al., 2017; Chen, Li, & Lei, 2012;Liu,Zhang,&Wang, 2010;Yu&Pan, 2008), but ignore the economic consequences of party-corporate relations. Some scholars use the ‘two-way access’ between CPCPO committee and board of directors in SOEs to study CHINA JOURNAL OF ACCOUNTING STUDIES 273 how CPCPO affects corporategovernanceand corporateperformance of SOEs (Chen& Lu, 2014;Maetal., 2012, 2013). However, only a few scholars have paid attention to CPCPO in Non-SOEs. Li, Meng, Wang, and Zhou (2008) and Dai, Yu, Ning, and Pan (2017) found that party membership of private business owners led to more loans, more judicial trust and fewer violations. Some scholars studied the economic beha- viours caused by CPCPO of Non-SOEs based on the National Private Enterprise Sampling Survey Data in 2006. Liang et al. (2010) found that Non-SOEs with CPCPO have more charitable donations. Long and Yang (2014) found that CPCPO can enhance employee welfare, and thus improve corporate performance. However, these researches only use a dummy variable to proxy whether the Non-SOEs had established CPCPO. This vari- able may not capture CPCPO’sreal effectiveness. Our study will comprehensively and deeply examine the relation between CPCPO and earnings management by using the number of activities organised by CPCPO to proxy the different power of CPCPO. 2.2. Hypothesis development Earnings management has always been an important field of corporate governance. The basic motivation for executives to be involved in the earnings management, is to unethi- cally obtain private gains from control (Shleifer & Vishny, 1997; Zingales, 1994). Specifically, this motivation is embodied in the form of increasing perquisite consump- tion, improving reputation and achieving the performance goals. Earnings management would make outsiders misjudge the performance of the firm, and affect the earnings- based contracts (Healy & Wahlen, 1999). It is the deprivation of the interests of share- holders, creditors and other stakeholders. In the long run, earnings management will reduce the corporate value (Xie, 2001). From the perspective of corporate governance, scholars have made rich studies on how to restrain earnings management. But under the background of a socialist market economy with Chinese characteristics, few scholars pay attention to the role of CPCPO in corporate governance. In China, the Non-SOEs usually are controlled by controlling shareholders. As a result, the traditional governance mechanisms, such as independent directors and institutional investors, have limited effectiveness on protecting minority shareholders’ interests. Consequently, it is important and interesting to ask what effects does CPCPO have on earnings management behaviour. 2.2.1. CPC party organisation in Non-SOEs According to the ‘embedding theory’ in social network, firms’ economic behaviours are always embedded in non-economic behaviour such as religion, culture, government and political parties (Granovetter, 1985). Non-economic behaviours affect economic beha- viours by embeddedness, thus bringing in governance effect (Uzzi, 1997). Hence, we are going to analyse how CPCPO affects Non-SOEs’ earnings management based on the embedding theory. From the micro view, CPCPO embeds into the Non-SOEs as a ‘formal institution’. Unlike the informal institution of fellow-villager and alumnus, the Company Law and the Party Constitution clearly stipulate that Non-SOEs should set-up CPCPO, and the authoritative force of CPCPO is supported by the CPC. CPCPO in Non-SOEs has formal members, funds and offices for running party organisation activities. CPCPO needs to hold regular meetings 274 D. ZHENG ET AL. and organises activities, such as leading the labour unions to organise social loving and public welfare activities, and learning the spirit of the 19th National Congress. Dai Yanjun, a professor of Central Party School of the CPC, pointed out that the key point of CPCPO in Non-SOEs is to make business owners realise how important CPCPO is. Therefore, CPCPO should actively invite the controlling shareholders, chairmen, CEOs and other senior execu- tives to participate in the party activities and meetings, and increase communication with them, which i would n turn make CPCPO have a greater influence in the Non-SOEs. In some Non-SOEs, their chairmen or CEOs is the secretary of the CPCPO committee. CPCPO in such Non-SOEs is more likely to be deeply embedded in their economic activities. In summary, at the micro-level, CPCPO can interact with employees through party organisation activities, actively attract the attention of chairmen and CEOs, and then can have real effects on the enterprise’s operation activities and corporate reporting behaviours. From the view of macro level, as a node, CPCPO of a single Non-SOE is embedded in the national network of CPCPO. Once a privately controlled enterprise has set-up a CPCPO, the CPCPO is like a ‘nerve ending’ of CPC because CPCPO needs to fulfil its tasks, report regularly to its leading higher party organisation, and is under the supervision of the higher party organisation. The national network of CPCPO and the network of enterprises have become an integrated dual network. Besides, the newly-presented network of CPCPOs also includes CPCPO alliance within a certain region. For example, Dayi in Sichuan Province sets up the first county-level Non-SOEs CPCPO alliance. This alliance has made remarkable achievements in promoting the development of enterprises and strengthening the construction of the party organisation in Non-SOEs. It is also very welcomed by the business owners and employees. Such regional alliances or national network of CPCPO bring the social credentials to business owners and enhance their identity as a party member for business owners (Lin, 2002). A strong network of CPCPO is conducive to play a better role in micro-enterprises. 2.2.2. The influence of CPCPO and earnings management As mentioned above, CPCPO embeds into the Non-SOEs and the latter also embeds into the national network of CPCPOs, which enables the non-economic behaviours of CPC to have a real effect on the economic behaviours of Non-SOEs. Then, how does CPCPO affect the earnings management behaviour of Non-SOEs? First, after CPCPO embeds into one Non-SOE, the CPCPO itself will play its active role in the organisation and actively implement its aims. One of the aims of CPCPO is to safe- guard the interests of all stakeholders which naturally includes the interest of investors, but earnings management does damage the investors’ interest. According to the provi- sions of the Party Constitution, CPCPO in Non-SOEs should implement CPC’s principles and policies, guide and supervise Non-SOEs and their management to abide by national laws and regulations. In order to safeguard the legitimate rights and interests of all parties and to promote the healthy development of enterprises, CPCPO needs to urge Non-SOEs to operate legally, and promote enterprises to balance the interests of insiders and other stakeholders. Therefore, under the guidance of the CPC’s aim of ‘serving the people’, CPCPO in Non-SOEs has sufficient impetus to urge Non-SOEs to safeguard the legitimate interests of all stakeholders and to abide by the laws and regulations, which may help to reduce the level of earnings management. Second, in Non-SOEs, the chairmen (they are usually the controlling shareholders) and CEOs are the ultimate decision makers. In order to achieve its goals, CPCPO will first actively CHINA JOURNAL OF ACCOUNTING STUDIES 275 get the recognition and support by the chairman and CEO after it embeds into one Non- SOE. If the chairman and CEO of one Non-SOE are communist party members, their identity as communist party member may be weakened because they have been away from the activities of CPCPO for a long time before their firm set-up CPCPO. Besides, the CPC’s spirit and advanced culture may also not be followed up in time before the firm set-up CPCPO. After the embedding of CPCPO, the chairman and CEO will return to CPCPO through a series of activities organised by CPCPO and constantly learning the new spirit of the CPC Central Committee. In addition, according to the ‘Upper Echelons Theory’,the CPC’s advanced culture will penetrate the chairman and CEO’s cognitive structure and their value system, which would deepen their sense of identity as a communist party member. Thus, the goals of CPCPO, ‘safeguarding the interests of all parties’, would be successfully implemented. What is more, even if the chairman and CEO of Non-SOEs are not communist party members, the firm should provide necessary conditions for the activities of CPCPO in accordance with the Party Constitution and the Company Law. And the 18th National Congress of CPC and other important national political conferences continued to emphasise the importance of CPCPO in Non-SOEs, which may attract more and more attention from the chairman and CEO of Non-SOEs. On the other hand, the CPCPO in the firm would invite the chairman and CEO to join some activities of CPCPO. Consequently, the chairman and CEO of the firm would actively participate the activities of CPCPO, and even apply to join the CPCPO. In the process of participating in the activities of CPCPO, the purpose and spirit of the CPC would deepen their minds and thus implicitly influence governance decisions such as earn- ings management (DiMaggio, 1997;Liu, 2016). For example, CPCPO of Zhejiang Baoye Construction Group Co., Ltd. has a goal to make the best managers become communist party members and also make the best communist party members become the best man- agers. In the company, 125 communist party members have been promoted to important management positions, and 42 excellent front-line talents have been recommended to being communist party members. Therefore, no matter whether the chairman and CEO of Non-SOEs are communist party members or not, they will be subtly influenced by the advanced culture and important spirit of the CPC as the result of the activities organised by CPCPO, which is helpful to restrain earnings management and other agency problems which do violate the disciplines and values pursued by the CPC. However, it cannot be ignored that, due to the limitation of activity conditions and the lack of proper understanding of CPCPO, CPCPO in Non-SOEs may be ‘nameless’, and even be used as a ‘rent-seeking tool’ by the Non-SOEs owners and insiders. To be specific, on the one hand, CPCPOs in Non-SOEs have to face insufficient manpower and material resources till now. Due to the late start of party construction in Non-SOEs, many grass- roots party officials do not fully understand the role of CPCPO in Non-SOEs In addition, as the result of the misunderstanding of the embedding of CPCPO and the financial con- straints of Non-SOEs, many Non-SOE enterprises do hesitate to set-up CPCPO and estab- lish CPCPO just to comply with the requirements of the Party Constitution and the Company Law Thus, the dual absence of subjective cognition and objective conditions make it difficult for CPCPO to play a positive role in Non-SOEs. On the other hand, the real Yang Chunlin, deputy director of municipal organisation department in Baoshan, Yunnan province, wrote an article on the official website of ‘Party Building of CPC in Private Enterprises’ and said that many private enterprises do not fully understand the CPCPO’s political core role among workers and the political leading role in production and operation. They are just ‘build for building’s sake’.. 276 D. ZHENG ET AL. purpose of the controlling shareholders of Non-SOEs is to use the establishment of CPCPO as a ‘bridge’ to get a political connection with local governments to acquire economic resources (Luo & Liu, 2009; Pfeffer & Salancik, 1978). CPCPO is only a kind of ‘rent-seeking tool’ and a ‘facade’ to whitewashing its illegal behaviours. If so, CPCPO not only could not play a substantial positive role in Non-SOEs, instead strengthens the motivation of the controlling shareholders and management to conduct earnings management to satisfy their private benefits (Liu, Zhou, Fu, & Xiao, 2010; Shen, Yang, & Pan, 2014; Tao, Cao, & Li, 2018). Under the above arguments, CPCPO may aggravate the earnings management behaviours of Non-SOEs. In short, the role of CPCPO of Non-SOEs in restraining earnings management is an open question waiting for empirical tests, we hypothesise that: Ceteris paribus, the CPCPOs have no effect on earnings management of Non-SOEs. 3. Research design 3.1. Regression specification In order to test the effect of CPCPO on earnings management, our regression specification is as following: jj DA ¼ β þ β PartyEffect þ β Controls þ Industry fe þ Year fe þ ε (1) 0 1 i The independent variable is CPCPO’seffect (PartyEffect), which is measured by CPCPO’s activities. Specifically, we include the following two variables: (1) PartyDum, a dummy variable set to 1 if the Non-SOE company has CPCPO activities in the year t, zero otherwise; (2) PartyNum, the degree of CPCPO’seffect (equal to the log value of the number of CPCPO activities in year t). The more CPCPO activities, the greater influence of CPCPO is. The reason why we use CPCPO activities to measure CPCPO’seffect is that CPCPO activity is the embodiment of its function, which reflects the vitality of CPCPO. The dependent variable is earnings management, including the following two vari- ables: (1) |DA_Adj| is the absolute value of the residual estimated from the modified Jones model (Dechow, Sloan, & Sweeney, 1995); (2) |DA_Kothari| the absolute value of the residual estimated from the performance-matched model (Kothari, Leone, & Wasley, 2005). The main control variables in model (1) include: 1. The corporate fundamental characteristics: current ratio, accounts receivable ratio, inventory ratio, tunnelling, return on net assets, cash flow from operating activities, loss, leverage, and size; 2. Corporate governance characteristics: the controlling shareholder’s shareholding ratio, whether the CEO is also the chairman. Finally, we control the industry and year-fixed the effect. The specific variable definitions are shown in Table 1. 3.2. Data source and sample selection Due to the availability of data, we use Non-SOE listed companies as our research sample. The sample range is from 2004 to 2015. We collect CPCPO activity data through the The secretary of the party committee in Yanan, Shaanxi province, wrote an article on the official website of ‘Party Building of CPC in Private Enterprises’ and said that party building in private enterprises started late and was a lake of funds, making it difficult to organise activities.. CHINA JOURNAL OF ACCOUNTING STUDIES 277 Table 1. Variables definition. Variables Definition |DA_Adj| Absolute value of the residuals from the modified Jones model (Dechow et al., 1995). |DA_Kothari| Absolute value of the residuals from the performance-matched model in Kothari et al. (2005). PartyDum An indicator equal to one for firm’s CPCPO organised party organisation activities in year t, equal to zero otherwise. PartyNum Log value of the number of activities organised by CPCPO in year t. The value indicates the party organisation’ influence. CR Current ratio, equal to the current assets divided by the current liability at the end of year t. AR Accounts receivable, equal to the accounts receivable scaled by total assets at the end of year t. INV Inventory ratio, equal to the book value of inventory scaled by total assets at the end of year t. Tunnelling Other receivables scaled by total assets at the end of year t. ROE Net profit scaled by average net assets in year t. OCF Cash flow from operating activities scaled by total assets at the end of year t. LOSS An indicator set to 1 if the net income is negative in year t, and zero otherwise. LEV The total assets divided by total liability at the end of year t. SIZE Log of total assets at the end of year t. FShr The shareholding ratio of the largest shareholder at the end of year t. Board Number of board of directors at the end of year t. Dual An indicator set to 1 if the chairman is also CEO, and zero otherwise. Out Independent directors percentage at the end of year t. AGE The number of years since firm IPO at the end of year t. HBShr An indicator set to 1 if the firm cross-listed in H share or B share, and zero otherwise. Big10 An indicator set to 1 if the auditor is a Big 10 auditor, and zero otherwise. LawIndex The degree of legalisation, equal to FanGang index sorted by 0–9 divided 9. The maximum value is 1 and the minimum value is 0. The higher the value is, the higher the degree of legalisation is. company’s home page, Baidu, and Google. The keywords for searching are ‘party’, ‘party committee’, ‘party organization’, ‘party branch’, ‘party construction’, ‘party member’, ‘communist youth league committee’, ‘communist youth league committee branch’, ‘communist youth league member’ and ‘labor unions’ (according to the relevant regula- tions, CPCPO leads the work of the Communist Youth League and Labour Unions). Then, we read those texts one by one and collect CPCPO activity data. CPCPO activity needs to meet the following two criteria: first, there were communist party members of the enterprise to participate the activity; second, the activity should be organised by CPCPO. An activity meets the two criteria above is recognised as CPCPO activity. After excluding the companies in the financial industry, ST companies and observations with missing data, we obtain 8,746 firm-year observations. The marketisation index comes from the ‘China Marketization Index’ and the other data is from the CSMAR database. We also cross-check those data with the WIND database and RESSET database. To control the influence of outliers, we winsorise all continuous variables at 1% and 99% level. 4. Results 4.1. Descriptive statistics Table 2-1 shows the descriptive statistics of the key variables used in our model. The mean of PartyDum and PartyNum are 0.29 and 0.44, respectively. The maximum and median value of PartyNum is 2.77 and 0.00, respectively. There is a big difference in the influence of party organisation among different firms. Regarding our dependent variables, the mean (median) of |DA_Adj| and |DA_Kothari| is 0.09 (0.06) and 0.08 (0.06). Other variables are basically the same as previous literature. 278 D. ZHENG ET AL. Table 2–1. Descriptive statistics for the full sample. Variable N Mean Std p1 Median p99 |DA_Adj| 8746 0.0876 0.0879 0.0010 0.0604 0.4440 |DA_Kothari| 7503 0.0825 0.0818 0.0010 0.0568 0.4170 PartyDum 8746 0.2910 0.4540 0.0000 0.0000 1.0000 PartyNum 8746 0.4400 0.7760 0.0000 0.0000 2.7730 CR 8746 2.3830 1.7780 0.4540 1.7010 6.5860 AR 8746 0.1210 0.0940 0.0028 0.1040 0.3390 INV 8746 0.1550 0.1220 0.0050 0.1250 0.4750 Tunnelling 8746 0.0241 0.0371 0.0009 0.0102 0.1790 ROE 8746 0.0731 0.0949 −0.1930 0.0719 0.2660 OCF 8746 0.0415 0.0713 −0.0942 0.0412 0.1860 LOSS 8746 0.0965 0.2950 0.0000 0.0000 1.0000 LEV 8746 0.4300 0.2990 0.0408 0.4060 1.5670 SIZE 8746 21.4000 0.9650 19.7700 21.3100 23.8600 FShr 8746 0.3230 0.1400 0.0903 0.2970 0.7080 Board 8746 8.5050 1.5970 5.0000 9.0000 14.0000 Dual 8746 0.3040 0.4600 0.0000 0.0000 1.0000 Out 8746 0.3710 0.0526 0.2860 0.3330 0.5560 AGE 8746 7.0500 5.3590 0.0000 5.0000 18.0000 HBShr 8746 0.0258 0.1590 0.0000 0.0000 1.0000 Big10 8746 0.4630 0.4990 0.0000 0.0000 1.0000 LawIndex 8746 0.6110 0.4880 0.0000 1.0000 1.0000 We further separate the sample into two subsamples by PartyDum. Table 2-2 shows the difference of the two subsamples. Compared with observations without CPCPO activities, observations with CPCPO activities show a lower level of earnings management. This evidence supports the hypothesis that CPCPO can restrain earnings management. In addi- tion, Table 2-2 also shows that there is no significant difference in performance and cash flow between the two subsamples. Observations with CPCPO activities have a larger firm size, a higher probability of hiring top 10 auditors, and have a lower leverage, a lower shareholding ratio of the largest shareholder, a lower ratio of independent directors and Table 2-2. Difference test. PartyDum =0 PartyDum=1 Variable N Mean Median N Mean Median T-test Wilcoxon-test |DA_Adj| 6199 0.0913 0.0620 2547 0.0787 0.0558 6.14*** 9.51*** |DA_Kothari| 5414 0.0859 0.0592 2089 0.0738 0.0521 5.76*** 15.47*** CR 6199 2.4230 1.7160 2547 2.2850 1.6760 3.30*** 2.06 AR 6199 0.1180 0.1010 2547 0.1270 0.1110 −3.76*** 11.64*** INV 6199 0.1540 0.1240 2547 0.1580 0.1300 −1.42 4.58** Tunnelling 6199 0.0265 0.0100 2547 0.0180 0.0100 9.76*** 5.21** ROE 6199 0.0734 0.0710 2547 0.0723 0.0720 0.45 0.34 OCF 6199 0.0421 0.0410 2547 0.0401 0.0410 1.18 0.09 LOSS 6199 0.1060 0.0000 2547 0.0746 0.0000 4.45*** 19.77*** LEV 6199 0.4340 0.4070 2547 0.4210 0.4050 1.94* 0.09 SIZE 6199 21.2900 21.1800 2547 21.6700 21.6100 −16.87*** 220.55*** FShr 6199 0.3270 0.2980 2547 0.3130 0.2910 4.37*** 2.61 Board 6199 8.4310 9.0000 2547 8.6860 9.0000 −6.79*** 24.10*** Dual 6199 0.3120 0.0000 2547 0.2860 0.0000 2.35** 5.52** Out 6199 0.3720 0.3330 2547 0.3680 0.3330 2.56** 2.28 AGE 6199 7.0440 5.0000 2547 7.0640 5.0000 −0.15 1.68 HBShr 6199 0.0260 0.0000 2547 0.0255 0.0000 0.12 0.01 Big10 6199 0.4210 0.0000 2547 0.5660 1.0000 −12.46*** 152.76*** LawIndex 6199 0.6120 1.0000 2547 0.6090 1.0000 0.27 0.24 Note: T-test is for mean difference test and Wilcoxon test is for median different test. ***, **, and* indicate 0.01, 0.05 and 0.10 significance levels, respectively. CHINA JOURNAL OF ACCOUNTING STUDIES 279 a lower probability of Chairman-CEO duality. Obviously, there are some differences in the company’s characteristics between these two subsamples. In order to control the systematic differences between the two subsamples, we use Propensity Score Matching (PSM) in the robustness test (after PSM matching, there is no significant difference in company’s char- acteristics of these two subsamples). Also, we use instrumental variables and DID model to controls the endogenous problems caused by differences in company characteristics in section 4.5. 4.2. Main results Table 3 reports the results of equation (1). The results show that PartyDum and PartyNum are both significantly and negatively correlated with earnings management (|DA_Adj| or | DA_Kothari|) at a significant level of 1%. With regard to the economic significance, when PartyDum increases one standard deviation, |DA_Adj| will decrease 3.40% (|DA_Kothari| will decrease 2.40%). When PartyNum increases one standard deviation, |DA_Adj| will decrease 3.40% (|DA_Kothari| will decrease 3.60%). These results suggest that CPCPO in Non-SOEs has a significant restraining effect on earnings management. 4.3. Cross-sectional analysis In this section, we examine how individual characteristics of chairman and CEO (i.e. whether they are communist party members, and whether they have political connec- tions), political sensitivity period and corporate-level characteristics (i.e. agency cost, information environment, financing constraints) affect the relationship between CPCPO and earnings management. We set the following variables: (1) Individual characteristics of chairman and CEO: an indicator for whether they are party member (ParMem). ParMem equal to 1 if the chairman or CEO is a party member, and zero otherwise; (2) An indicator variable for the politically sensitive period (NPC). Generally speaking, political sensitivity is higher during National People’s Congress of the CPC. NPC equal to 1 if the year is 2007 or 2012, and zero otherwise. (3) Characteristics at the company level: Wedge, an indicator equal to 1 if the separation of cash flow right and control right of ultimate controller is greater than the 75 quantiles (higher agent costs), and zero otherwise; Syn, an indicator equal to 1 if stock price synchronicity is greater than the 75 quantiles (lower information environment) , and zero otherwise (Piotroski & Roulstone, 2004); KZ, an indicator equal to 1 if the company’sKZ financing constraint index is greater than the 75 quantiles (higher financing constraint) , and zero otherwise (Kaplan & Zingales, 1997). Tables 4 and 5 reports the results of the relationship among CPCPO, individual characteristics of chairman and CEO and earnings management. We find that CPCPO and earnings management are still significantly and negatively correlated. ParMem and earnings management is also significantly and negatively correlated, and PolCon is positively correlated with earnings management but not significant. More importantly, we find that the coefficient of the interaction term between CPCPO and ParMem is We do not report the results of control variables since Table 4 for brevity. We use R2 of CAPM model to capture stock price synchronicity. And it is calculated by weekly stock return and weekly market return. KZ index is calculated according to the model of Kaplan and Zingales (1997). 280 D. ZHENG ET AL. Table 3. The effect of CPCPO on earnings management. (1) (2) (3) (4) Variable |DA_Adj| |DA_Kothari| |DA_Adj| |DA_Kothari| PartyDum −0.0057*** −0.0054*** (−2.86) (−2.69) PartyNum −0.0037*** −0.0036*** (−3.13) (−3.04) CR 0.0034*** 0.0022*** 0.0034*** 0.0022*** (4.00) (2.66) (4.01) (2.67) AR −0.0534*** −0.0557*** −0.0529*** −0.0552*** (−4.17) (−4.42) (−4.14) (−4.38) INV −0.0287*** −0.0256** −0.0289*** −0.0258** (−2.62) (−2.30) (−2.64) (−2.31) Tunnelling 0.1724*** 0.1394*** 0.1722*** 0.1392*** (4.55) (3.64) (4.54) (3.64) ROE 0.0827*** 0.0785*** 0.0830*** 0.0787*** (5.33) (5.26) (5.35) (5.28) OCF −0.1602*** −0.1525*** −0.1596*** −0.1519*** (−8.47) (−8.08) (−8.45) (−8.06) LOSS 0.0111** −0.0009 0.0112*** −0.0008 (2.57) (−0.20) (2.60) (−0.18) LEV 0.0453*** 0.0351*** 0.0453*** 0.0351*** (7.00) (5.80) (7.01) (5.81) SIZE −0.0037*** −0.0020 −0.0037*** −0.0020 (−3.03) (−1.61) (−3.00) (−1.58) FShr 0.0163** 0.0226*** 0.0166** 0.0228*** (2.24) (3.05) (2.27) (3.08) Board −0.0013* −0.0008 −0.0013* −0.0008 (−1.93) (−1.23) (−1.89) (−1.20) Dual 0.0032 0.0017 0.0032 0.0016 (1.58) (0.83) (1.55) (0.78) Out −0.0253 −0.0112 −0.0251 −0.0111 (−1.29) (−0.56) (−1.28) (−0.56) AGE 0.0005* 0.0005* 0.0005** 0.0005** (1.88) (1.94) (1.97) (2.02) HBShr −0.0065 −0.0069 −0.0065 −0.0068 (−0.99) (−1.02) (−0.98) (−1.01) Big10 −0.0041** −0.0019 −0.0041** −0.0019 (−2.03) (−0.94) (−2.04) (−0.94) LawIndex −0.0037* −0.0050** −0.0037* −0.0051** (−1.78) (−2.38) (−1.82) (−2.42) Intercept 0.1489*** 0.1130*** 0.1477*** 0.1118*** (5.13) (3.86) (5.08) (3.82) Year & Ind FE Include Include Include Include N 8746 7503 8746 7503 Adj-R2 0.0860 0.0825 0.0861 0.0828 Note: T value is in the parentheses and is clustered by firm; ***, **, and * indicate 0.01, 0.05 and 0.10 significance levels, respectively. significantly negative, and the coefficient of the interaction term between CPCPO and PolCon is also significantly negative. These results show that the negative relationship between CPCPO and earnings management is more pronounced in the companies where the chairman or CEO is a party member or they have a political connection. Table 6 show the regression results of CPCPO, political sensitivity period and earnings management. We find that CPCPO and earnings management are still significantly and negatively correlated. Relationship between NPC and earnings management are significantly negative. In addition, the interaction term between CPCPO and political sensitivity period is CHINA JOURNAL OF ACCOUNTING STUDIES 281 Table 4. CPCPO, communist party member and earnings management. |DA_Adj| |DA_Kothari| Variable (1) (2) (3) (4) PartyDum −0.0049** −0.0052** (−2.15) (−2.34) PartyDum×ParMem −0.0035*** −0.0030** (−2.79) (−2.16) PartyNum −0.0027** −0.0030** (−1.99) (−2.24) PartyNum×ParMem −0.0035** −0.0022* (−2.41) (−1.87) ParMem −0.0016*** −0.0008** −0.0002** −0.0007** (−2.58) (−2.31) (−2.09) (−2.28) Controls Include Include Include Include Year & Ind FE Include Include Include Include N 8746 8746 7503 7503 Adj-R2 0.0859 0.0862 0.0823 0.0826 Note: T value is in the parentheses and is clustered by firm; ***, **, and * indicate 0.01, 0.05 and 0.10 significance levels, respectively. Table 5. CPCPO, political connections and earnings management. |DA_Adj| |DA_Kothari| Variable (1) (2) (3) (4) PartyDum −0.0056*** −0.0053** (−2.75) (−2.57) PartyDum×PolCon −0.0027** −0.0038** (−2.30) (−2.39) PartyNum −0.0034*** −0.0033*** (−2.83) (−2.76) PartyNum×PolCon −0.0073*** −0.0072*** (−3.38) (−3.24) PolCon 0.0017 0.0004 0.0002 0.0018 (0.33) (0.09) (0.03) (0.34) Controls Include Include Include Include Year & Ind FE Include Include Include Include N 8746 8746 7503 7503 Adj-R2 0.0856 0.0859 0.0823 0.0827 Note: T value is in the parentheses and is clustered by firm; ***, **, and * indicate 0.01, 0.05 and 0.10 significance levels, respectively. significantly negative, which shows that the negative relationship between CPCPO and earn- ings management is more significant during political sensitive period. Tables 7–9 show the regression results of CPCPO, company characteristics and earnings management. We find that CPCPO and earnings management are still significantly and negatively correlated. Wedge is positively correlated with earnings management, Syn is negatively correlated with earnings management, and KZ is positively correlated with earnings management. In addition, we find that the inter- action terms between CPCPO and the three variables are all significantly negative. These results show that when the company’s agency costs is higher, the information environment is much worse and the financing constraint is stronger, CPCPO has 282 D. ZHENG ET AL. Table 6. CPCPO, political sensitivity period and earnings management. |DA_Adj| |DA_Kothari| Variable (1) (2) (3) (4) PartyDum −0.0071*** −0.0066*** (−3.16) (−2.85) PartyDum×NPC −0.0065** −0.0045** (−2.38) (−2.01) PartyNum −0.0048*** −0.0046*** (−3.70) (−3.39) PartyNum×NPC −0.0051* −0.0037** (−1.80) (−2.40) NPC −0.0163* −0.0160* −0.0123 −0.0125 (−1.71) (−1.67) (−1.54) (−1.56) Controls Include Include Include Include Year & Ind FE Include Include Include Include N 8746 8746 7503 7503 Adj-R2 0.0860 0.0863 0.0825 0.0828 Note: T value is in the parentheses and is clustered by firm; ***, **, and * indicate 0.01, 0.05 and 0.10 significance levels, respectively. Table 7. CPCPO, agency cost and earnings management. |DA_Adj| |DA_Kothari| Variable (1) (2) (3) (4) PartyDum −0.0059*** −0.0049** (−2.59) (−2.12) PartyDum×Wedge −0.0072*** −0.0059** (−2.69) (−2.24) PartyNum −0.0039*** −0.0036*** (−2.92) (−2.64) PartyNum×Wedge −0.0073*** −0.0063** (−2.84) (−2.47) Wedge 0.0011 0.0010 0.0015 0.0001 (0.26) (0.38) (0.36) (0.03) Controls Include Include Include Include Year & Ind FE Include Include Include Include N 8746 8746 7503 7503 Adj-R2 0.0867 0.0868 0.0833 0.0835 Note: T value is in the parentheses and is clustered by the firm; ***, **, and * indicate 0.01, 0.05 and 0.10 significance levels, respectively. astrongerrestraining effect on earnings management, which highlights the active governance effect of CPCPO in Non-SOEs. 4.4. Additional analyses In addition to the above results, we first further test the potential mechanism that CPCPO affects earnings management. Whether CPCPO has real governance effect in Non-SOEs highly depends on how the chairman or CEO treat CPCPO. If the chairman or CEO actively participate in CPCPO activities, or even serve as party secretary, it is more likely that CPCPO plays a greater governance role Non-SOEs in the company. Thus, CPCPO could affect the company’s governance decision-making and financial behaviour. We use two CHINA JOURNAL OF ACCOUNTING STUDIES 283 Table 8. CPCPO, information environment and earnings management. |DA_Adj| |DA_Kothari| Variable (1) (2) (3) (4) PartyDum −0.0033** −0.0048** (−2.33) (−2.03) PartyDum×Syn −0.0061*** −0.0072** (−2.58) (−2.46) PartyNum −0.0027* −0.0038*** (−1.87) (−2.73) PartyNum×Syn −0.0043** −0.0038** (−2.06) (−2.35) Syn 0.0059** 0.0066*** 0.0082*** 0.0091*** (5.13) (5.09) (3.85) (3.82) Controls Include Include Include Include Year & Ind FE Include Include Include Include N 8746 8746 7503 7503 Adj-R2 0.0869 0.0870 0.0841 0.0843 Note: T value is in the parentheses and is clustered by firm; ***, **, and * indicate 0.01, 0.05 and 0.10 significance levels, respectively. Table 9. CPCPO, financial constraints and earnings management. |DA_Adj| |DA_Kothari| Variable (1) (2) (3) (4) PartyDum −0.0029** −0.0033** (−2.54) (−2.32) PartyDum×KZ −0.0122*** −0.0147*** (−2.90) (−2.92) PartyNum −0.0034** −0.0037*** (−2.56) (−3.00) PartyNum×KZ −0.0076*** −0.0096*** (−3.12) (−3.42) KZ 0.0096*** 0.0093*** 0.0097*** 0.0096*** (2.91) (2.88) (2.90) (2.95) Controls Include Include Include Include Year & Ind FE Include Include Include Include N 8746 8746 7503 7503 Adj-R2 0.0872 0.0875 0.0841 0.0845 Note: T value is in the parentheses and is clustered by firm; ***, **, and * indicate 0.01, 0.05 and 0.10 significance levels, respectively. indicator variables to proxy the attitude of the chairman or CEO: (1) Partake, an indicator equal to 1 if the chairman or CEO participates CPCPO activities, and zero otherwise; (2) Secretary, an indicator equal to 1 if the chairman or CEO is the CPCPO secretary, and zero otherwise. The results are shown in Table 10. Partake and Secretary are significantly and negatively correlated with earnings management at least at the level of 5%, which implies that the participation of chairman and/or CEO in the CPCPO activities is very important and might play as the channel. In addition, the audit committee is mainly responsible to supervise a firm’s financial statement disclosure and internal control process, and aims to improve accounting information quality. Then, if a party member of CPC also sit on the audit committee, it would also be a potential channel that CPCPO affects earnings management. We then set- 284 D. ZHENG ET AL. Table 10. Mechanism test. |DA_Adj| |DA_Kothari| Variable (1) (2) (3) (4) (5) (6) Partake −0.0101*** −0.0086** (−2.74) (−2.33) Secretary −0.0087*** −0.0081*** (−3.87) (−3.61) Party_Shenji −0.012* −0.015** (−1.89) (−2.08) Controls Include Include Include Include Include Include Year & Ind FE Include Include Include Include Include Include N 8746 8746 8746 7503 7503 7503 Adj-R2 0.0859 0.0865 0.0907 0.0823 0.0830 0.0878 Note: T value is in the parentheses and is clustered by firm; ***, **, and * indicate 0.01, 0.05 and 0.10 significance levels, respectively. up a variable to proxy CPC party member ratio of the audit committee, Party_Shenji, which is equal to the number of party members of CPC in audit committee, divided by the total numbers of audit committee members. The results are also shown in Table 10. Party_Shenji is also significantly and negatively correlated with earnings management at least at the level of 10%, which implies that the higher the ratio of CPC party member ratio in audit committee, the better accounting quality is provided. Second, we use two variables to distinguish the substance of CPCPO activities: Party_Theory, an indicator equal to 1 if CPCPO activities involve theoretical study activities (i.e. learning the Party Constitution and CPC’s conference spirit), and zero otherwise. Party_Practice, an indicator equal to 1 if CPCPO activities involve practical activities (i.e. visiting revolutionary bases, unpaid blood donations and other kinds of donations). The regression results are shown in Table 11.Both Party_Theory and Party_Practice are significantly and negatively correlated with earnings management. The coefficient of Party_Practice is relatively larger, but there is no significant differ- ence between the two coefficients. These results indicate that both CPCPO’stheore- tical study activities and practical activities do help to restrain earnings management. Third, we try to rule out the following two possible alternative explanations: (1) Earnings management does not necessarily have a negative effect on company value (Fang & Fu, 2018). Earnings management can convey private information and Table 11. Theoretical study activities and practice activities. |DA_Adj| |DA_Kothari| Variable (1) (2) Party_Theory −0.0025* −0.0031* (−1.90) (−1.75) Party_ Practice −0.0042** −0.0049** (−2.15) (−2.04) Difference test F = 0.10, P = 0.76 F = 0.50, P = 0.47 Controls Include Include Year & Ind FE Include Include N 8746 7503 Adj-R2 0.0910 0.0886 Note: T value is in the parentheses and is clustered by firm; ***, **, and * indicate 0.01, 0.05 and 0.10 significance levels, respectively. CHINA JOURNAL OF ACCOUNTING STUDIES 285 thus correct undervalued stock prices, and it is more likely to pass on private information through earnings management in undervalued companies. The embed- ding of CPCPO strengthens the communication among CPCPO, chairman and CEO and other stakeholders, and then reduce insiders’ need to transfer private informa- tion through earnings management. If so, CPCPO could reduce the demand for earnings management’s information transmission role, then the negative relationship between CPCPO and earnings management should be more pronounced in the companies where the stock price is more likely to be underestimated. Thus, we divide our sample into two subsamples based on PEG, which is an indicator equal to 1 if P/E divided by growth rate of net profit over the next 3 years is between 0 and 1, and zero otherwise. PEG between 0 and 1 means the stock price is under- valued. The coefficient between PEG and earnings management is about 0.05 and significant at 1%, which means it is possible to provide private information through earnings management in undervalued companies. Panel A of Table 12 shows the results based on the two subsamples: The relationship between CPCPO and earnings management is significantly negative in the non-undervalued group, but the nega- tive correlation is not significant in the undervalued group, which help us exclude the above alternative explanation. (2) The negative results between CPCPO and earnings management may not be driven by CPCPO’s active governance role that urging Non-SOEs to safeguard the interests of all parties, but only because CPCPO makes a company feel pressure to behave more carefully. As a result, the company with active CPCPO would reduce earnings management to makes the company unobtrusively. If so, the negative correla- tion between CPCPO and earnings management will be more pronounced in the ‘more cautious’ companies. We divide our sample into two subsamples based on whether the enterprise would get sanction in the next 3 years for the financial fraud in year t. Companies would be more cautious after doing fraud behaviours in year t.The results are shown in Panel B of Table 12. The results show that the negative correlation between CPCPO and earnings management is significantly negative in the non-sanction group, but not significant in the sanction group. In short, we also exclude this alternative interpretation. 4.5. Robustness test We conduct the following robustness tests First, we use two-stage least square method (2SLS) to alleviate the potential endogen- ous problems such as omitted variable bias. Considering that Non-SOEs with better corporate governance are more inclined to build CPCPO, our results may be endogenous. We then use two instrumental variables PolValue and AiguozhuyiBase to alleviate the endogenous problems PolValue is from the World Values Survey in 2007 and 2012 on the political values of citizens in all provinces of China Besides, we manually collected data to set the other instrumental variable, AiguozhuyiBase, which is equal to one if a city has In order to control the government intervention brought by CPCPO, we also include Fangang government intervention index as a control variable. The result is still robust. We do not report the result for brevity.. We also use a dynamic GMM model with two lag periods of dependent variables to control endogeneity. The results are still robust. We do not report the results for brevity.. 286 D. ZHENG ET AL. Table 12. Excluding alternative explanations. |DA_Adj| |DA_Kothari| |DA_Adj| |DA_Kothari| (1) (2) (3) (4) (5) (6) (7) (8) under non-under under non-under under non-under under non-under Variable valued valued valued valued valued valued valued valued Panel A: Whether the stock price is undervalued PartyDum -0.0061 -0.0052** -0.0044 -0.0056** (-1.31) (-2.38) (-0.97) (-2.52) PartyNum -0.0010 -0.0043*** -0.0002 -0.0045*** (-0.34) (-3.44) (-0.09) (-3.51) Controls Include Include Include Include Include Include Include Include Year & Ind Include Include Include Include Include Include Include Include FE N 2093 6653 1873 5630 2093 6653 1873 5630 Adj-R2 0.0789 0.0924 0.0764 0.0910 0.0782 0.0930 0.0760 0.0917 |DA_Adj| |DA_Kothari| |DA_Adj| |DA_Kothari| (1) (2) (3) (4) (5) (6) (7) (8) Variable sanction non- sanction sanction non- sanction sanction non- sanction sanction non- sanction Panel B: Whether the firm gets sanction in the next 3 years PartyDum -0.0036 -0.0050** -0.0035 -0.0047** (-1.07) (-2.40) (-1.12) (-2.20) PartyNum -0.0011 -0.0036*** -0.0012 -0.0035*** (-0.21) (-2.91) (-0.25) (-2.85) Controls Include Include Include Include Include Include Include Include Year & Ind FE Include Include Include Include Include Include Include Include N 628 8118 628 6875 628 8118 628 6875 Adj-R2 0.0604 0.0880 0.0596 0.0856 0.0588 0.0882 0.0579 0.0860 Note: T value is in the parentheses and is clustered by firm; ***, **, and * indicate 0.01, 0.05 and 0.10 significance levels, respectively. the ‘National Patriotism Education Demonstration Base’ in year t. The ‘National Patriotism Education Demonstration Base’ is set-up by the propaganda department of CPC Central Committee. The regression results of 2SLS are shown in Panel A of Table 13. According to the first stage regression results listed in column 1–2, the F values are 23.959 and 29.583, respec- tively, which means that the two instrumental variables are not weak instrumental vari- ables. In addition, the Hansen J statistic shows that our IVs are valid instruments, i.e. uncorrelated with the error term. What is more, according to the second stage results of 2SLS in columns 3–6, we find that the results are consistent with the main results in Table 3. Second, we use PSM method to control the systematic differences between the observations with and without CPCPO activities. In the first stage of PSM, besides all the control variables in Table 4, we added four additional variables that directly affect CPCPO activities: ParMem, whether the chairman or CEO is a CPC member or not; PolCon, whether the chairman or CEO has political connection or not; NPC, whether it is the political sensitive period or not; GaiZhi, whether the firm is privatised from SOE or not. The regression results are shown in Table 13 Panel B. Using the PSM matching sample, we The political values section in the World Values Survey includes three sub-items: the degree of political importance, the degree of interest in politics and the degree of trust in political party. The smaller the original value is, the stronger the political consciousness is. PolValue is equal to the mean of the three sub-items multiplied by −1, and then the higher the value of PolValue, the stronger the political consciousness. We use the political values data in 2007 to capture PolValue before 2012, and the data in 2012 to capture PolValue in other years.. CHINA JOURNAL OF ACCOUNTING STUDIES 287 Table 13. Robustness test. First stage of the 2SLS model Second stage of the 2SLS model PartyDum PartyNum |DA_Adj| |DA_Kothari| Variable (1) (2) (3) (4) (5) (6) Panel A: Instrumental variable AiguozhuyiBase 0.0305*** 0.0728*** (3.07) (4.38) PolValue 0.0583*** 0.1034*** (6.19) (6.30) PartyDum -0.0648** -0.0902*** (-2.42) (-3.41) PartyNum -0.0343** -0.0471*** (-2.49) (-3.45) Controls Include Include Include Include Include Include Year & Ind FE Include Include Include Include Include Include N 8746 8746 8746 8746 7503 7503 Adj-R2 0.1363 0.1306 0.0858 0.0858 0.0832 0.0832 F Value of Weak IV test 23.9590*** 29.5830*** Hansen J statistic (p value) 0.6196 0.7839 |DA_Adj| |DA_Kothari| Variable (1) (2) (3) (4) Panel B: PSM PartyDum -0.0053** -0.0043* (-2.38) (-1.90) PartyNum -0.0035*** -0.0032** (-2.82) (-2.48) Controls Include Include Include Include Year & Ind FE Include Include Include Include N 5094 5094 4171 4171 Adj-R2 0.0886 0.0890 0.0922 0.0927 |DA_Adj| |DA_Kothari| Variable (1) (2) Panel C: PSM+DID Post 0.000 0.002 (0.11) (0.46) Treat×Post -0.008*** -0.009*** (-2.82) (-3.32) Controls Include Include Year & Firm FE Include Include N 5945 5147 Adj-R2 0.1415 0.1512 |DA_DD| Opacity Variable (1) (2) (3) (4) Panel D: Alternative independent variables PartyDum -0.0031* -0.0224*** (-1.65) (-4.68) PartyNum -0.0029*** -0.0122*** (-2.65) (-4.59) Controls Include Include Include Include Year & Ind FE Include Include Include Include N 6550 6550 5812 5812 Adj-R2 0.06 0.06 0.17 0.17 |DA_Adj| |DA_Kothari| Variable (1) (2) (3) (4) Panel E: Control other factors PartyDum -0.0057*** -0.0054*** (-2.87) (-2.69) PartyNum -0.0037*** -0.0036*** (Continued) 288 D. ZHENG ET AL. Table 13. (Continued). |DA_Adj| |DA_Kothari| Variable (1) (2) (3) (4) (-3.14) (-3.05) PolCon 0.0075* 0.0066* 0.0059 0.0080 (1.89) (1.72) (1.20) (1.22) GaiZhi 0.0065** 0.0067** 0.0054* 0.0056** (2.30) (2.38) (1.92) (2.00) Other Controls Include Include Include Include Year & Ind FE Include Include Include Include N 8745 8745 7503 7503 Adj-R2 0.0910 0.0912 0.0886 0.0888 Note: T value is in the parentheses and is clustered by firm; ***, **, and * indicate 0.01, 0.05 and 0.10 significance levels, respectively. find that CPCPO is still significantly negatively correlated with earnings management, which is consistent with the main regression results. Third, we use PSM-DID method to control the endogenous problem. We use firms that never organise CPCPO activities as a control sample (Treat=0), and use firms that have organised CPCPO activities as treat group (Treat= 1). We match the treat group with a control group using one-to-one PSM method in the first year of organising CPCPO activity. Second, we set-up an indicator variable Post, which equal to 1 after the first year of organising CPCPO activity, and zero before the first year (the value of Post of control firm is the same as its matched treat firm). The results are shown in Table 13 Panel C. After controlling firm and year fixed effect (Treat is automatically omitted due to firm-fixed effect), we find that the coefficient of Treat×Post is significantly negative, suggesting that, com- pared with companies that have no CPCPO activities, earnings management of companies with CPCPO activities are significantly lower after the first year of organising CPCPO activity. Fourth, we also use other measurement for dependent variables to do robust test. Besides |DA_Adj| and |DA_Kothari|, we use DD model (Dechow & Dichev, 2002) to measure earnings management (|DA_DD|). The results show that CPCPO is still significantly and negatively correlated with earnings management. Also, we use Opacity to measure information opacity. Opacity is calculated as the sum of |DA_Adj| in year t-3 to year t-1. Results in Table 13 Panel D show that CPCPO is also significantly and negatively correlated with information opacity, which are consistent with the main regression results. Last, we also add PolCon and GaiZhi as control variables, because these two factors may affect both CPCPO activeness and earnings management (Liu et al., 2010;Shenetal., 2014; Tao et al., 2018; Wang & Liu, 2014). The results are showed in Table 13 Panel E, and our conclusions remain. 5. Conclusions The construction of CPCPO in Non-SOEs has achieved remarkable progress since the 18th National Congress of CPC and the implement of important documents such as the Opinions. But there arelittleacademicresearches on whether and how CPCPO in Non-SOEs play real governance role. This paper uses the manually collected data of A-shares Non-SOE listed companies from 2004 to 2015 to investigate whether and how CPCPO helps to restrain the earnings management behaviour of Non-SOE firms. We document that the relationship CHINA JOURNAL OF ACCOUNTING STUDIES 289 between CPCPO and earnings management is significantly negative, and the negative relationship is more pronounced when the chairman or CEO is CPC member, has political connections, and when it is the political sensitivity period. Besides, this relationship is also more pronounced in companies with higher agency costs, poorer information environment, and higher financing constraints. The possible mechanisms are to encourage chairmen and CEOs of those firms to participate in the activities of CPCPO and even serve as the secretaries, and to make directors whom are CPC party member enter audit committee. After controlling possible endogenous problems and excluding alternative interpretations, the above results remain. These results imply that CPCPO could play real governance role and restrain earnings management, and the greater the influence of CPCPO, the better the governance role. Based on the above findings, we put forward the following suggestions: Firstly, for the CPC workers in Non-SOEs, the first task of CPCPO is to win the support of the top management such as the chairman and CEO. Because the empirical results of this paper show that the governance effectiveness of CPCPO is stronger in those firms whose chairmen and CEOs are more involved in the activities of CPCPO and chairmen or CEOs serve as the party secretary. Secondly, for the regulators of capital market, they should more actively promote the construction of CPCPO in Non-SOE firms. Compared with developed countries, the history of China’s capital is too short and premature. Under the background of socialism with Chinese characteristics, the embedding of CPCPO into Non-SOEs is an effective corporate governance mechanism, which can restrain the earnings management of Non-SOEs to some extent. Thirdly, for the top-level departments of CPC who are in charge of the party construc- tion works, it is important to strengthen the cohesion of CPCPO in Non-SOEs network and strengthen the social credit proof function of CPCPO network. For example, the top-level departments of CPC should give reward to those companies whose CPCPO play better governance role. Thus, the construction of CPCPO in Non-SOEs will be more and more acceptable by the controlling shareholders, and chairman and CEOs are more likely to support CPCPO, which would help CPCPO to achieve its goal better. Disclosure statement No potential conflict of interest was reported by the authors. Funding This work was supported by the National Natural Science Foundation of China (71672098 and 71802206), the Humanity and Social Science Youth foundation of Ministry of Education (18YJC630262), and Fundamental Research Funds for the Central Universities (QL18005). References Chen, D.Q., Chen, Y.S., & Dong, Z.Y. (2017). Policy uncertainty, market competition and allocation of capital: Evidence from turnover of city-level leaders. Journal of Financial Research, 11,65–80. In Chinese. Chen, D.Q., Li, S.F., & Lei, G.Y. (2012). The influence of Local government governance and control structure over the investment – evidence from family firm. Journal of Financial Research, 3, 124–138. In Chinese. 290 D. ZHENG ET AL. Chen, S.H., & Lu, C.R. (2014). Can the party organization in SOE effectively curb the “loss of state- owned assets” in mergers and acquisitions? Management World, 5, 106–120. In Chinese. Dai, Y.Y., Yu, W., Ning, B., & Pan, Y. (2017). Chairman’s communist party member status and financial violations of private enterprises. Accounting Research, 6,75–81. In Chinese. Dechow, P.M., & Dichev, I.D. (2002). The quality of accruals and earnings: The role of accrual estimation errors. The Accounting Review, 77,35–59. Dechow, P.M., Sloan, R.G., & Sweeney, A.P. (1995). Detecting earnings management. The Accounting Review, 70, 193–225. DiMaggio, P. (1997). Culture and cognition. Annual Review of Sociology, 23, 263–287. Fang, V.W., & Fu, R.H. (2018). The bright side of earnings management. (Working paper). Retrieved from https://ssrn.com/abstract=3224800 Fisman, R. (2001). Estimating the value of political connections. The American Economic Review, 91, 1095–1102. Goldman, E., Rocholl, J., & So, J. (2009). Do politically connected boards affect firm value? The Review of Financial Studies, 22, 2331–2360. Granovetter, M. (1985). Economic action and social structure: The problem of embeddedness. American Journal of Sociology, 91, 481–510. Healy, P.M., & Wahlen, J.M. (1999). A review of the earnings management literature and its implica- tions for standard setting. Accounting Horizons, 13, 365–383. Julio, B., & Yook, Y. (2012). Political uncertainty and corporate investment cycles. The Journal of Finance, 67,45–83. Kaplan, S., & Zingales, L. (1997). Do investment cash flow sensitives provide useful measures of financing constraints? Quarterly Journal of Economics, 112, 169–215. Kothari, S.P., Leone, A.J., & Wasley, C.E. (2005). Performance matched discretionary accrual measures. Journal of Accounting and Economics, 39, 163–197. Li, H., Meng, L., Wang, Q., & Zhou, L. (2008). Political connections, financing and firm performance: Evidence from Chinese private firms. Journal of Development Economics, 87, 283–299. Liang, J., Chen, S.Y., & Gai, Q.G. (2010). Political participation, governance structure and charitable donation of private enterprises. Management World, 7, 109–118. In Chinese. Lin, N. (2002). Social capital: A theory of social structure and action. Cambridge: Cambridge University Press. Liu, H.L., Zhang, M., & Wang, Y.P. (2010). Political connections, compensation incentive, and employee allocation efficiency. Economic Research Journal, 9, 109–121. In Chinese. Liu, Q.L., Zhou, L.H., Fu, J., & Xiao, J. (2010). Political connection, private relationship, choice of audit firm and audit conspiracy. Auditing Research, 4,66–77. In Chinese. Liu, X. (2016). Corruption culture and corporate misconduct. Journal of Financial Economics, 122, 307–327. Long, X.N., & Yang, J. (2014). CCP committee worker benefits and firm performance: Evidence from Chinese private firms. China Journal of Economics, 2, 150–169. In Chinese. Luo, D.L., & Liu, X.L. (2009). Political relations, barriers to entry and corporate performance – Evidence from China’s private listed companies. Management World, 5,97–106. In Chinese. Ma, L.F., Wang, Y.F., & Shen, X.X. (2012). Research on governance effects of China’s state-owned companies’ party organization——A perspective based on “insiders control”. China Industrial Economics, 8,82–95. In Chinese. Ma, L.F., Wang, Y.F., & Shen, X.X. (2013). Party organization governance of SOE, redundant employ- ees and executive compensation contracts. Management World, 5, 100–115. In Chinese. Pfeffer, J., & Salancik, G. (1978). The external control of organizations. New York: Harper and Row. Piotroski, J.D., & Roulstone, D.T. (2004). The influence of analysts, institutional investors, and insiders on the incorporation of market, industry, and firm-specific information into stock prices. The Accounting Review, 79, 1119–1151. Shen, H.B., Yang, Y.L., & Pan, F. (2014). Political connection, securities violation and earnings quality of private listed companies. Journal of Financial Research, 1, 194–206. In Chinese. Shleifer, A., & Vishny, R.W. (1997). A survey of corporate governance. The Journal of Finance, 52, 737–783. CHINA JOURNAL OF ACCOUNTING STUDIES 291 Tao, X.H., Cao, S.W., & Li, C.T. (2018). Political connections and earnings management: Evidence from China. Review of Industrial Economics, 4,30–44. In Chinese. Uzzi, B. (1997). Social structure and competition in inter-firm networks: The paradox of embeddedness. Administrative Science Quarterly, 42,35–67. Wang, K.M., & Liu, B. (2014). Corporate control transfer and earnings management. Management World, 7, 144–156. In Chinese. Xie, H. (2001). The mispricing of abnormal accruals. The Accounting Review, 76, 357–373. Yu, M.G., & Pan, H.B. (2008). Political connection, institutional environment and private enterprise bank loan. Management World, 8,9–21. In Chinese. Zingales, L. (1994). The value of the voting right: A study of the milan stock exchange experience. Review of Financial Studies, 7, 125–148.
Journal
China Journal of Accounting Studies
– Taylor & Francis
Published: Apr 3, 2019
Keywords: Communist party of China; party organisation; privately-controlled enterprises; earnings management
