The power of belief: party organization construction of accounting firms and audit quality

ABSTRACT We examine the impact of party organisation construction of accounting firms on audit quality. We find that although there is no relationship between the party membership status of auditors and audit quality, the audit quality of party-member auditors significantly improves after accounting firms strengthen party organisation construction. Additional analyses show that the positive impact of auditor party membership on audit quality does not generate a contagious effect in the presence of party organisation construction. Moreover, cross-section results show that the impact of party organisation construction on audit quality is more pronounced in Non-Big4 audit firms, and in audit firms having a higher proportion of party-member partners/employees. Finally, we find that party organisation construction of accounting firms plays a more significant role in improving audit quality in state-owned firms. We believe our paper to be the first to investigate the governance role of party organisation construction in accounting firms.


Introduction
Establishing party organisations in enterprises is a systemic arrangement with Chinese characteristics. The Constitution of the Communist Party of China and the Company Law of the People's Republic of China articulate requirements for party organisation construction of enterprises. In recent years, in order to facilitate the Party's core role of taking overall charge and coordinating all entities, the Party Central Committee has attached great importance to party organisation construction in new social organisations such as accounting firms. Since 2010, The CPA industry began comprehensively strengthening the construction of party organisations as advocated by the Ministry of Finance and other departments. As party organisation construction within accounting firms deepens, these organisations have begun to participate in internal governance, to the point where they now exert significant influence on the latter.
In contemporary academic circles, the governance role of party organisations has become a key research priority. Many scholars have focused on state-owned enterprises to study the effect of party organisations' participation in governance (Chen & Lu, 2014;Cheng et al., 2015;Hao & Ma, 2018;Huang et al., 2017;Lai, 2018;Ma et al., 2012Ma et al., , 2013Wu & Wang, 2018), while others have analysed the effect of such participation in private enterprise governance (Dong & Wei, 2018;He & Ma, 2018;Liang et al., 2010;Long & Yang, 2014;Ye, 2017). Most of this literature concludes that party organisations play a positive role in corporate governance. However, no one has yet studied the effect of party organisation participation in the governance of accounting firms.
Considering that: 1) accounting firms, which provide broad-spectrum financial activity oversight, and play irreplaceable roles in the capital market (and, arguably, in the national economy as a whole) (Francis, 2011;M. DeFond & Zhang, 2014); 2) the question of how accounting firm governance affects audit quality has become an important issue in recent years, attracting attention from all quarters (C. Wang et al., 2015); and 3) whether and how, as a form of participation in their internal governance, party organisation participation affects audit quality, are valuable questions to ask; to repeat, considering these three points our paper examines the effect of party organisation participation on accounting firms' governance on the specific issue of audit quality.
Political beliefs or ideologies have profound impacts on human behaviour (Dai et al., 2017). Previous studies show that beliefs lay the foundation for accountants' moral decision-making (Craig et al., 2007;Ding et al., 2012). Therefore, we first examine whether auditors' political beliefs affect audit quality. However, practice shows that long-term separation from party organisations and disengagement from party organisational life tends to undermine party members' political beliefs. Party organisation construction can be seen as a means to address these issues; such construction might also strengthen the political beliefs of party member auditors through participation in an accounting firm's internal governance. Therefore, we examine whether accounting firms' party organisation construction strengthens party member auditors' political beliefs, thereby leading to improvements in audit quality. Further, sociological studies maintain that patterns of human behaviour spread through social networks (Lawrence et al., 2011;M. Liu & Qiao, 2014). Accepted patterns of behaviour and values are likewise spread through various channels within accounting firms; studies on the contagious effects of audit quality provide sufficient empirical evidence for this assertion (Francis & Michas, 2013;Lawrence et al., 2011). Over an extended period, members of the Communist Party of China have participated in market activities, bringing with them an exemplary social image as pioneers. Therefore, whether the political beliefs of party member auditors have a contagious effect becomes an interesting question. In addition, considering that both auditor and customer characteristics are key factors affecting audit quality (Chen, 2014;Francis & Krishnan, 1999;Ireland & Lennox, 2002;Li & Luo, 2011), we believe it is worth investigating whether the heterogeneity of auditors and firms affects the relationship between party organisation construction and audit quality.
Using publicly-traded companies in Chinese A-Shared markets from 2004 to 2016 as our sample, we examine the impact of accounting firms' party organisation construction on audit quality. Results show that, although there is no significant difference in audit quality between party and non-party member auditors, the audit quality of party-member auditors significantly improves after accounting firms strengthen party organisation construction. This conclusion holds after a series of robust tests. Further analyses show that positive impacts of auditors' political beliefs on audit quality have no contagious effect under the background of the accounting firms' party organisation construction. Specifically, we find no improvement in audit quality for those non-party auditors who have cooperated with party auditors in the previous year after accounting firms strengthen party organisation construction. When further considering the heterogeneity of accounting firms, we find that compared with these Big4 firms, the positive impact of party organisation construction on audit quality is more pronounced for non-Big4 firms. Moreover, we contend that, under the background of accounting firms' party organisation construction, the higher the proportion of party member employees/partners in accounting firms, the greater the audit quality. Finally, when considering differences in clients' property rights, we see that accounting firms' party organisation construction plays a more significant role in audit quality improvements for those state-owned enterprises.
Our paper contributes to the literature in several ways. First, previous studies have investigated the effects of accounting firms' internal governance on audit quality with respect to accounting firm governance of branches, information construction, and information disclosure (Deumes et al., 2012;C. Zeng et al., 2018;C. Wang et al., 2015). By contrast with these studies, we investigate the impact of accounting firms' internal governance on audit quality from the perspective of party organisation participation in governance. In this way, our paper expands the research framework of accounting firms' internal governance, and enriches the literature on the impact of accounting-firm internal governance on audit quality.
Third, previous studies have generally investigated the influence of auditors' individual attributes on audit quality (DeFond & Zhang, 2014;Gul et al., 2013;Li et al., 2017;M.L. DeFond & Francis, 2005). We focus on a more specific or granular dimension of auditors' individual characteristics, namely, their individual political beliefs. In so doing, we add a novel perspective to the literature on the impact of individual auditor attributes or inclinations on audit quality.
Fourth, our research offers empirical evidence on whether and how an accounting firm's party organisation construction affects audit quality, thus lending empirical support for party organisation construction in accounting firms.
Our paper is organised as follows. Section 2 introduces the institutional background, reviews the related literature, and develops hypotheses. Section 3 introduces the main variables, model design, sample, and data. Section 4 presents the descriptive statistics and empirical results. Section 5 covers the additional tests, and Section 6 serves as conclusion.

Institutional background
In recent years, the CPA industry in China has expanded, recognition of its professional role has grown, and its social impact has markedly increased. To ensure healthy and stable capital market development and to protect the public interest, it is thus important to promote party organisation construction in accounting firms and other organisations. In February 2008 at the National Organizational Work Conference, Xi Jinping stated that we should establish and strengthen the foundations of party organisation construction in social organisations. He went on to note that, because they employ large workforces and have specific regulatory departments, accounting and law firms constitute a sort of threshold point. More importantly, they have laid a solid foundation for further work. Following his statement, accounting firms began to further explore the idea. By mid-October 2009, the Party Group of the Central Committee of the CPC, along with the CPC Ministry of Finance, issued a notice on Further Strengthening the Party Organization Construction in the CPA industry (hereafter referred to as 'Notice'). The Notice formed the basis for policy and theoretical guidance, which helped enable accounting firms to conduct party organisation construction work; it was a genuine milestone in the party organisation construction of the Chinese CPA industry. In addition, the CPA industry Party Committee was established on 18 October 2009. These developments came to constitute a 'backbone' for the development of the CPA industry, laying the foundations for comprehensively strengthening party organisation construction work therein.
Under the guidance of the Notice and the Party Committee, the CPA industry in China has established a management system of party construction. The system combines all elements, and gives full play to the role of the party organisation. In so doing, it realises the 'double coverage' of party organisation and work, and the 'double rationalization' of accounting firms' party organisation subordinate relationships on one hand, and party member organisation relationships on the other. Moreover, through their active promotion of this new regime, accounting firms have recorded outstanding achievements in party organisation construction. According to a report in the People's Daily, before October 2009 there were 1,032 party organisations in Chinese accounting firms. By the end of 2010, this number had reached 3,352, a total representing full coverage of party organisations and party work across the industry (see Figure 1 for details).
In addition, party organisation construction in accounting firms also pays heed to daily business, and carries out themed annual activities and internal governance. This generates widespread impacts that manifest in the improvement of firm' internal governance.

Research on the effect of formal systems on audit quality
With the rapid development of the Chinese CPA industry, governance of accounting firms has encountered various problems. To address these, numerous policies and systems have been introduced with a view to standardising governance and ensuring audit quality; in turn, these have attracted much scholarly attention. Extant studies have examined the impact of formal systems (such as mandatory audit rotation) on audit quality (Gong & Wang, 2009;Lennox et al., 2014;Zhang et al., 2011;H. Xu et al., 2017). Similar research has addressed restricting employment relationships between accounting firms and clients (Chen et al., 2015;Geiger et al., 2008Geiger et al., , 2005Lennox, 2005); changing accounting firms' organisational form (Chen et al., 2016;Zhang & Wei, 2016); and other external supervision systems (Liu, 2016;Song, 2011;Wu, 2008;Zhu & Wu, 2009). The bulk of this literature concludes that formal systems are important as mechanisms to ensure audit quality. However, to date there has been no investigation of the impact of formal systems on audit quality from the specific perspective of party organisation construction.

Research on the effect of culture system on audit quality
Accounting firm culture is the sum of those beliefs and values that influence auditors' behaviour, as well as a value system with which auditors within a given firm agree, to the extent that their behaviour is constrained by that system. Compared with formal mechanisms such as organisational structure, rules, and regulations, culture has a more farreaching impact on auditor behaviour (Jenkins et al., 2008;Neimark & Tinker, 1986). Moreover, appropriate corporate culture is a useful supplement to formal governance mechanisms; it plays an important and positive role in improving accounting firms' internal governance and, in turn, the quality of their audits. Several studies have shown that accounting firm culture affects audit quality (Carpenter et al., 1994;Windsor & Ashkanasy, 1996). However, these studies mainly explore the impact of culture on audit quality at the level of the firm.
Notably, as an important element of culture personal belief is now a salient issue in the field of culture, so that numerous studies have examined the influence of beliefs on individual behaviour. For example, scholars have studied the impact of executive beliefs on corporate behaviour, finding that the religious beliefs of executives may affect corporate risk preference (Jiang et al., 2015;Miller & Hoffmann, 1995); decision-making (Hilary & Hui, 2009;Lei et al., 2016); corporate irregularities (Chen et al., 2013); and charitable donations (Zeng et al., 2016). In addition, researchers have noted that political beliefs might also influence individual behaviour (Dai et al., 2017;Ding et al., 2012;Gul et al.,  2013). Some scholars have queried the effects of political belief on auditor behaviour, albeit without reaching consistent conclusions (Ding et al., 2012;Gul et al., 2013). Examples here include Ding et al. (2012), who suggest that there is no relationship between auditors' political beliefs and audit quality; on the other hand, Gul et al. (2013) find that audit quality among party-member auditors is lower. We investigate the effects of auditors' political beliefs on audit quality based on the background of party organisation construction in accounting firms.

Hypotheses
Belief arguably influences behaviour in ways that are imperceptible, as in deeply embedded patterns of cognition. Similarly, faith could lay an unseen foundation for accountants to make moral decisions (Craig et al., 2007;Ding et al., 2012). The influence of communist thought on party members is reflected in their values and thinking patterns (Zhou & Hu, 2014). We argue similarly that communist belief exerts an influence on auditors' values and behaviour, thereby adding to audit quality.
First, communism contains a unique value concept, which steers party-member auditors towards sound value judgements. Article 2 of the Constitution of the Communist Party of China points out that party members always take 'serving the people wholeheartedly' as a basic premise. In this way, the spirit of altruism is from the outset deeply entrenched in these auditors' modes of thinking, helping them to determine normatively preferential courses of action. Moreover, the ingrained attitude of serving the people dovetails nicely with the public good that naturally inheres in the auditing enterprise. In other words, steeped as they are in communist beliefs, party-member auditors are better able to carry out their mission of serving the people and the furthering public interest. In addition, the Constitution of the Communist Party points out that communist party members should follow the concept of seeking truth from facts; this serves as an additional guide for member auditors when forming their judgements. In terms of specific practices, engaging in, say, collusion, would work directly against investors' interests not to mention being diametrically opposed to the spirit of altruism, or to the principle of seeking truth from facts. Under the guidance of unique value concepts, it is less likely that party-member auditors would collude or otherwise engage in compromising activity with clients; rather, the impetus would be towards conducting high quality audits. Therefore, compared with non-party member auditors, party member peers are less likely to tolerate, say, clients' earnings manipulations; instead, we might observe a greater tendency for party member auditors to issue modified audit opinions or audit reports issued by party member auditors are less aggressive.
Second, the image of exemplary vanguard encourages party member auditors to constantly improve audit quality. The Communist Party of China is characterised by a tradition of nobility and excellence, bathed in the light of brilliant deeds by revolutionary martyrs and outstanding party members, and it continues each day to bring this spirit forward. Communist Party members have for decades and in their every action carried the banner of exemplary pioneer; indeed, membership itself stands as a signal of competence and integrity, even as these qualities typically remain unobservable, including in member auditors (Dai et al., 2017;Liu & Wang, 2010). To embody and maintain this identity as exemplar, party-member auditors have a strong, built-in motivation to improve audit quality. On the one hand, they guard their independence, thus avoiding participation in audit malfeasance. On the other, they seek to constantly maintain and improve a high level of professional competence in their work, striving to reduce the possibility of errors in judgement, or other faults that might adversely affect an audit. In short, the image of exemplary pioneer serves as enduring encouragement to party-member auditors, making it much more likely for them to refuse to tolerate clients' manipulation of earnings, or to allow instances of fraud and conversely more likely to issue modified audit opinions, and to ensure overall quality.
Finally, strict party rules and discipline tend to favourably shape party-member auditors' behaviour. Compared with their non-party peers, member auditors not only abide by basic market laws and regulations, but are also bound by party discipline and regulations. For example, Article 33 of the Regulations on Disciplinary Measures of the Communist Party of China clearly states that if a party member violates laws, rules, or regulations, disciplinary sanctions will result. Because of these guiding constraints, party members can be expected to scrupulously follow the Auditing Standards for Chinese CPAs, thus demonstrating prudence in their practices. Given the above, we propose the following hypothesis:

H1a: Ceteris paribus, party-member auditors perform higher quality audits than nonparty member auditors.
A contrary position has been expounded: some studies view party membership as a proxy for political connections or political capital (Huang, 2013;Gul et al., 2013). Individuals with connections or capital tend to use such leverage to improve their own positions. Moreover, political connections may also constitute a sort of protective umbrella, helping tempted individuals to avoid regulatory and penalty risks and thus impairing the efficacy of regulatory enforcement (Chen & Lu, 2014;Shen et al., 2014;N. Xu et al., 2013). According to this argument, party membership provides cover for auditors who might then benefit from reduced supervision and punishment risk. In turn, this may lead to deteriorations in audit quality as compared to non-party member audits (Gul et al., 2013).
Furthermore, party members may be affected by a variety of other factors. In particular, the ideology of long-term inactive party members may stray from party doctrine in deleterious ways. Party members caught in anti-corruption campaigns since the 18th CPC National Congress demonstrate that the fundamental reason for corrupt behaviour is a lack of party spirit. Worth noting here is an article from the website of the Ministry of Supervision of the Central Discipline Commission titled "One of the Situation and Tasks in the Construction of Party Conduct and Clean Government and the Fight against Corruption: the Situation Determines the Task", in which the root cause of spreading corruption is variously identified as the indifference of some leading cadres to the party's ideology, lax organisation, and lax discipline. The characteristics of CPA industry make auditors more likely to flow, and the composition of members become more complicated. Some party members have been divorced from their original units for a long time, thus they have not joined party organisation lives for many years. As a result, some party member auditors' ideological concepts have been seriously divorced from their organisations and the people, which may lead to the shaking of party member auditors' political beliefs, thus eliminating the positive impact of auditors' political beliefs on audit quality. In this light, we offer the following hypothesis:

H1b: Ceteris paribus, there is no significant difference in audit quality between partymember auditors and non-party member auditors.
As stated above, the Notice on Further Strengthening Party Construction in CPA Industry has promulgated requirements for Party organisation construction in accounting firms. Such construction is helpful in strengthening party-member auditors' political beliefs, and in the ultimate improvement of audit quality.
First, the initiative will strengthen the sense of organisational belonging in mobile party members, allowing them to 'return' to the organisation, which in turn addresses the problem of alienation. Second, given this backdrop party committee secretaries at all levels will come to be regarded as 'key minorities' by CPA industry party committees and party committees will implement leadership responsibility mechanisms within the various organisations. It should be noted that committee secretaries of party organisations are selected and appointed from party leaders or party-member partners of accounting firms. In particular, it is advised that the secretary of the party organisation and the chief accountant should be the one and the same individual. We know that the chief accountant or partner has great influence on decision-making, and on accounting firm management and governance. Thus, where and whenever the chief accountant or the chief partner is also head of party affairs, then patterns of behaviour and values as manifest in party members will be embedded in the internal governance activities of the chief accountant or partner. This should have a desirable effect on both party-and nonparty member auditors.
Second, party organisation construction promotes honesty within accounting firms: it requires not only that party-member partners-and -auditors stand as examples in terms of honesty and principles of independence, but also that they encourage partymember partners-and -auditors to take the lead in acquiring new technologies according to their professional knowledge levels. In this way, auditors' professional competence can be improved.
Finally, accounting firms should strengthen the acquisition of talent by vigorously promoting accounting firms in cultivating and absorbing qualified elites and solid professionals into party organisation. These measures might not only strengthen party ranks, but also put those CPAs demonstrating generous party spirit and outstanding professional ability into key positions. Obviously, these measures give full play to the vanguard and exemplary role of party-member auditors.
In a word, we argue that accounting firms' party organisation construction can foster the altruistic spirit of party-member auditors to serve the people, enhance the exemplary pioneer spirit, and strengthen discipline and compliance with the law. Taken together, these effects will help buttress party-member auditors' political beliefs, and improve audit quality. Therefore, we propose the following hypothesis:

Audit quality measures
According to DeFond and Zhang (2014), audit quality can be measured via audit process outputs. We use multiple measures to proxy for audit quality, including accrual quality (AbsDA), audit reporting aggressiveness (ARAgg) and accounting conservatism (C_Score) (Dechow et al., 2010;Gul et al., 2013;Gong et al., 2016). First, we use accruals quality (AbsDA) as proxies for audit quality, which is calculated by using the accrual estimation error approach developed by Kothari et al. (2005). Second, following prior studies (e.g. DeFond et al., 2000;Francis & Krishnan, 1999;Gul et al., 2013), we use the predicted probability of issuing modified audit opinions minus the actual value to measure audit reporting aggressiveness (ARAgg). Third, we follow Dechow et al. (2010) and Gong et al. (2016), using accounting conservatism (C_Score) to measure audit quality.

Explanatory variables
Two types of explanatory variables are included in this paper. First, we use a dummy variable PARTY to measure party-member auditors. PARTY equals 1 if one of the two signing auditors' signatures has party member status, and 0 otherwise. Second, we use two dummy variables to measure accounting firms' party organisation construction. As mentioned earlier, at the National Organizational Work Conference in February 2008 Xi Jinping noted that we should explore and strengthen grass-roots party building within social organisations. Following this, accounting firms entered the exploratory stage of party organisation construction. Later (in mid-October 2009), the Party Group of the Central Committee of the CPC and the Ministry of Finance of the CPC jointly issued the Notice on Further Strengthening the Party Organization Construction in the CPA Industry; this became the basis of policy and theoretical guidance for CPA firms to conduct party organisation construction. The CPA Industry party committee was also established at this time, suggesting that CPA firms have comprehensively strengthened party organisation from this date onwards. As such, we use two dummy variables to capture the impact of CPA firms' party organisation construction on audit quality. POST1 captures the effect of CPA firms' party organisation construction in the exploratory stage, which equals 1 for the period 2008-2009, and 0 otherwise. POST2 captures the effect of CPA firms' party organisation construction in the comprehensive strengthening stage, which equals 1 for the period of 2010-2016, and 0 otherwise.

Control variables
As suggested in previous studies (Chan et al., 2006;Dechow et al., 2010;Wang et al., 2008), we include three sets of control variables that may influence audit quality. First, we include several variables that reflect clients' characteristics. Size equals the natural logarithm of total assets. ROA is measured as total profit divided by total assets. Leverage is the ratio of total liabilities to total assets. Growth equals firms' revenue growth ratio. Loss equals 1 if the client has reported a loss, and 0 otherwise. Quick is the ratio of current assets to current liabilities. Inv is the ratio of inventory divided by total assets. Rev is the ratio of net receivables to total assets. MTB equals market value divided by book value at year-end. Board equals the total number of a firm's directors. Inddep equals the ratio of independent directors. SOE equals 1 if the client is ultimately controlled by the government, and 0 otherwise.
Second, we control variables represented audit firms' characteristics. Big4 equals 1 if firms' auditors belong to Big Four international accounting firms, and 0 otherwise. Tenure is audit tenure. Change equals 1 if a firm's auditor changed in the current year, and 0 otherwise. Lnfee equals the natural logarithm of audit fees. Transfor equals 1 if a firm's auditor is a special general partnership, and 0 otherwise.
Third, we control variables captured auditors' characteristics. Impor1 equals a client's total assets divided by all clients' total assets of a review partner at one year. Impor2 equals a client's total assets divided by all clients' total assets of an engagement auditor at one year. Age1 equals the natural logarithm of a review partner's age. Age2 equals the natural logarithm of an engagement partner's age. Female1 equals 1 if a review partner is female, and 0 otherwise. Female2 equals 1 if an engagement partner is female, and 0 otherwise. Major1 equals 1 if a review partner majored in finance, accounting or auditing, and 0 otherwise. Major2 equals 1 if an engagement partner majored in finance, accounting or auditing, and 0 otherwise. Education1 equals 1 if a review partner at least holds a bachelor's degree, and 0 otherwise. Education2 equals 1 if an engagement partner at least holds a bachelor's degree, and 0 otherwise. Partner1 equals 1 if a review auditor is a partner, and 0 otherwise. Apartner2 equals 1 if an engagement auditor is a partner, and 0 otherwise. Experience1 equals the number of years a review partner has served as a signing partner of listed company. Experience2 equals the number of years an engagement partner has served as a signing partner of listed company.
In addition, we also controlled for firm fixed effects (Firm FE), year fixed effects (Year FE), and audit firm fixed effects (AuditFirm FE) .

Empirical model
In order to exclude unobservable factors that do not change with time, we adopt a fixedeffect model to test our hypotheses.
First, we establish the following model to test H1a and H1b: Where Audit Quality is proxied by three different measures, including accrual quality (AbsDA), audit reporting aggressiveness (ARAgg) and accounting conservatism (C_Score). PARTY equals 1 if one of the two signing auditors' signatures has party member status, and 0 otherwise. The coefficient on PARTY it reflects whether there is a significant difference in audit quality between party-member auditors and non-party member auditors.
Second, we use the following model to test H2: Where Audit Quality is proxied by three different measures, including accrual quality (AbsDA), audit reporting aggressiveness(ARAgg) and accounting conservatism (C_Score). PARTY equals 1 if one of the two signing auditors' signatures has party member status, and 0 otherwise. POST1 captures the effect of audit firms' party organisation construction in the exploratory stage, which equals 1 for the period 2008-2009, and 0 otherwise. POST2 captures the effect of audit firms' party organisation construction in the comprehensive strengthening stage, which equals 1 for the period of 2010-2016, and 0 otherwise. The coefficient on the interaction term PARTY it ×POST 1 reflects the effect of audit firms' party organisation construction in the exploratory stage on party-member auditors in improving audit quality. The coefficient on the interaction term PARTY it ×POST 2 reflects the effect of audit firms' party organisation construction in the comprehensive strengthening stage on party-member auditors in improving audit quality.

Sample and data
We conduct an empirical analysis of companies that were publicly traded in Chinese A-Shared markets from 2004 to 2016. We obtain data from several sources. First, financial and industry information on clients, and information on auditors and audit firms, derives from the China Stock Market and Accounting Research (CSMAR) database. Second, we manually collect information on the background of auditors (e.g. gender, age, education degree, and political identity) from the CICPA website. Third, the data on the proportion of party members in accounting firms from 2010 to 2012 is gleaned from CICPA. In addition, we exclude those companies with delisting risk warnings, companies belonged to financial and insurance industry, B-share (foreign share) firms and firms with missing data. Furthermore, to reduce the impact of uneven sample distribution, we removed companies that went public after 2008. Table 1 reports descriptive statistics of all variables. We winsorise all continuous variables at the 1st and 99th levels to mitigate the potential influence of outliers. The mean (median) value of AbsDA is 0.065 (0.044), and the standard deviation of AbsDA is 0.066, which together mean that the degree of earnings manipulation varies greatly among different enterprises. The mean (median) value of ARAgg is 0.022 (0.024); the standard deviation of same is 0.163, this indicates that audit reports aggressive varies greatly among different enterprises. The mean (median) value of C_Score is 0.043 (0.036), which is similar to Gong et al. (2016). The mean value of PARTY is 0.419, indicating that about 41.9% percent of the companies in the sample had signature auditors with party status. The mean value of POST1 is 0.137, indicating that roughly 13.7% of the sample is in the exploration stage of party organisation construction. The average value of POST2 is 0.654, such that 65.4% of the sample is in the stage of comprehensive strengthening its party organisation construction. The distribution of control variables is within a reasonable range. Table 2 reports the results of the relationship between party-member auditors and audit quality. In column (1), the coefficient on PARTY is negative but not significant, which indicates that party membership of auditors does not significantly improve client's   (3), the coefficient on PARTY is negative but not significant, suggesting that party membership of auditors does not significantly reduce audit report radicalisation. In column (5), the coefficient on PARTY is positive but not significant, indicating that auditor party membership does not significantly improve client's accounting conservatism. These results suggest that the party membership of auditors does not significantly improve audit quality. The reason that auditors with political beliefs do not achieve higher audit quality may be that political beliefs amongst party auditors have been shaken after longstanding disengagement from party organisation, and concomitant cessation of participatory activity. That is, our results support H1b. Based on this finding, we investigate the relationship between party membership of auditors and audit quality in the context of CPA-industry party organisation construction. The results are given in columns (2), (4), and (6) of Table 2. In column (2), the coefficient on PARTY×POST1 is −0.003 and insignificant (t-value = −0.785); but the coefficient on PARTY×POST2 is −0.006 and significant at 10% level (t-value = −1.895). In column (4), the coefficient on PARTY×POST1 is −0.011 and insignificant (t-value = −1.174); but the coefficient on PARTY×POST2 is −0.018 and significant at 5% level (t-value = −2.478). In column (6), the coefficient on PARTY×POST1 is 0.003 and insignificant (t-value = 1.376); but the coefficient on PARTY×POST2 is 0.005 and significant at 1% level (t-value = 2.688). These results suggest that party organisation construction of accounting firms at comprehensive strengthening stages is helpful in terms of audit quality improvement. In addition, we calculated the economic significance of the impact of accounting firms' party organisation construction on party-members' audit quality. Our results show that quality improvement as a result of party organisation construction in firms is at least at 9% level in terms of economic significance. Firstly, we calculate the economic significance of the impact of party organisation construction on clients' earnings quality for partymember auditors. We find that the value of AbsDA decreases by 9.23% (=0.006 × 1 × 1/ 0.065) on average at the comprehensive strengthening stage. Secondly, we calculate the economic significance of the impact of party organisation construction on accounting conservatism for party-member auditors. Our results show that the value of C_Score decreases by 11.63% (=0.005 × 1 × 1/0.043) on average at the comprehensive strengthening stage. Finally, we calculate the economic significance of the impact of party organisation construction on the degree of aggressive audit reports for party-member auditors. Our results show that the value of ARAgg decreases by 81.81% (=0.018 × 1 × 1/0.022) on average at the comprehensive strengthening stage.

Basic empirical results
The above results indicate that clients' earnings quality, accounting conservatism, and audit reports aggressive for party-member auditors are significantly improved both in statistical and economic senses as a consequence of accounting firms' party organisation construction. This implies that accounting firms' party organisation construction strengthens the political beliefs of party members, and ultimately spurs them to improve audit quality. Overall, our findings support H2.

PSM method
Although the regression results indicate that accounting firms' party organisation construction can strengthen political beliefs among party members and ultimately improve audit quality, our findings may suffer from an endogeneity problem. Specifically, differences exist between the clients of party-member auditors and those of their non-party peers. To account for this, we adopt the PSM method (Rosenbaum & Rubin, 1985) to obtain one-to-one matching observations. In particular, we use the nearest neighbour PSM method without replacement, controlling for all control variables in equation (1). Using treatment and control group samples, we repeat our regression and present the results in Table 3.
The coefficients of PARTY remain insignificant in columns (1), (3), and (5) of Table 3 after controlling for the potential endogeneity problem. This means that there is no significant difference in audit quality between party-member and non-party member auditors. The coefficients of PARTY×POST1 remain insignificant in columns (2) and (4), and the coefficients of PARTY×POST2 remain significant in columns (2), (4), and (6) of Table 3 after controlling for the potential endogeneity problem, from which we conclude that party organisation construction assists party-member auditors in raising audit quality. These results are consistent with our previous findings.

Self-selection
We examine the impact of accounting firm party organisation construction on partymember auditors' behaviour, finding that introducing this process strengthens belief amongst party-member auditors, which in turn leads to audit quality improvements. However, this conclusion may overlook the problem of sample self-selection. For example, it may be the case that party-member auditors prefer firms with higher earnings quality; conversely, firms (especially state-owned enterprises) may prefer either party-member auditors or accounting firms with a higher proportion of such auditors. To address this possibility, we use the Heckman two-stage method to solve the potential self-selection problem. In the first stage, we take PARTY as the dependent variable, and control for all control variables in equation (1); we also add an instrumental variable (Z). Next, we use the probit model to get the inverse mills ratio (IMR), and then introduce that variable in the second stage. The instrumental variable Z equals 1 if a company is registered in the old revolutionary base area, and 0 otherwise. We choose this variable based largely on the following considerations. First, Angrist and Krueger (2001) points out that the externality of instrumental variables can only be confirmed by economic theory and background. Therefore, we examine the exogenous variables from economic theory and background as the latter relate to instrumental variables. The old revolutionary base area has a strong communist ideology and culture. Listed companies in this area are more likely to be influenced by Communist culture, and are similarly more likely to be drawn towards party-member auditors or accounting firms whose proportion of party members are higher. In addition, the formation of the old revolutionary base area is not affected neither by the earnings quality of a given enterprise nor the identity of the auditor, and the fact that being in the old revolutionary base area has no direct impact on earnings quality allows us to consider it as a suitable instrumental variable.   4 reports the results of the Heckman two-stage procedure. In column (1), the coefficient of Z is 0.062, and is significant at 5% level (z-value = 2.471). This means that listed companies in the old revolutionary base area are more likely to choose partymember auditors, which is consistent with our prediction. In columns (2), (4), and (6), the coefficients on PARTY are insignificant, that is, there is no marked difference in audit quality between party-member and non-party member auditors after controlling for potential self-selection problems. In columns (3), (5) and (7), the coefficients on PARTY×POST1 are insignificant, but the coefficients on PARTY×POST2 are all significant, that is, after we control for potential self-selection problems, the audit quality of partymember auditors significantly improves after accounting firms strengthen party organisation construction. In other words, these results are consistent with our previous findings.

Placebo test
Although we take various steps to ensure robust results, other causal factors may play a role. To eliminate this possibility, we conduct a placebo test. Specifically, we take companies that were publicly traded in Chinese A-Shared markets from 2002 to 2014 as samples, and generate two time-dummy variables (Post1 and Post2) to capture the impact of accounting firm party organisation construction. Dummy variable Post1 equals 1 for the period 2006-2007; and 0 otherwise. Dummy variable Post2 equals 1 for the period 2008-2014; and 0 otherwise. Compared with samples period (2004-2016) as noted above, the sample period (2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014) for this section was advanced by two years. Moreover, compared with dummy variables POST1 and POST2 as noted above, Post1 and Post2 denote two years forward. If accounting firms' party organisation construction strengthen belief amongst party-member auditors (and triggers a concomitant improvement in audit quality), then this will appear in the coefficients of PARTY×post1 and PARTY×post2. Results are shown in Table 5. In columns (2), (4) and (6), the coefficients of PARTY×post1 are insignificant, as well as the coefficients of PARTY×post2 in columns (2) and (6), only the coefficient on PARTY×post2 in column (4) is significant at 5% level. That is, when compared with coefficient significance levels of PARTY×POST1 and PARTY×POST2 in Table 2, those of PARTY×post1 and PARTY×post2 in Table 5 are obviously reduced. These findings suggest that the changes as noted above are genuinely caused by audit firms' party organisation construction rather than by other factors.

Using the full sample
As previously mentioned, to reduce the impact on our conclusions of uneven sample distribution, we delete companies listed after 2008. To test for robustness here, we now reintroduce those companies; results are shown in Table 6. In columns (1), (3), and (5), the coefficients on PARTY are insignificant, while in columns (4) and (6), coefficients of POST1× PARTY and POST2× PARTY are similar to our previously reported results.

Contagious effect test
Previous studies argue for a contagious effect on accounting firms' audit quality (Francis & Michas, 2013;M. Liu & Qiao, 2014). To extrapolate, we query whether the influence of auditors' political beliefs on audit quality exerts similar influence. Sociological research shows that patterns of behaviour spread through social networks (Francis & Michas, 2013). In a social network, rational actors judge which behaviours are desirable based on direct observation or on language communication; they consequently internalise these desirable behaviours (M. Liu & Qiao, 2014). In other words, acceptable values and patterns of behaviour spread through accounting firms according to certain institutional arrangements or social networks. More specifically, junior members of the firm tend to observe and adopt senior-member behaviour, making it their own. In much the same way, members of equal standing also form mutually recognised standards through their communication. If so, then auditors' business models or conceptions of value are to a degree contagious within the firm. One might then expect party-member behaviour to positively if imperceptibly affect non-member colleagues, given the backdrop of a firm's party organisation construction.
In order to examine this potential effect, our sample was processed as follows: First, we identify non-party member auditors who have cooperated with their partymember peers. Second, we remove any observations that one of the two signing auditors' signatures has party member status. The remaining samples include nonparty auditors who have cooperated with party-member auditors in previous year as well as those who have not. Based on this, we generate a dummy variable CO_PARTY, which equals 1 if one of two non-party member auditors has cooperated with party-member colleagues in the previous year, and 0 otherwise. Then, we use CO_PARTY to replace PARTY, repeat our regression, and present the results in Table 7.
In columns (1), (3), and (5) of Table 7, the coefficients on CO_PARTY are insignificant, suggesting that compared with non-party member auditors who did not cooperate within the previous year, there is no difference in the audit quality of those non-party members who had cooperative experiences. In columns (2), (4), and (6), the coefficients on POST1× CO_PARTY and POST2× CO_PARTY are all insignificant, indicating that, compared with non-party auditors who had no cooperation experience with party-member peers, there is once again no difference in audit quality when measured against cooperating non-party members (under conditions of party organisation construction within accounting firms). We conclude that the positive impact of accounting firms' party organisation construction on audit quality is chiefly felt among party-member auditors. The pioneering role of party-member auditors has not exerted a contagious effect on non-party member auditors.

The effect of audit firm size
As stated above, accounting firms' party organisation construction arguably strengthens party-member auditors' political beliefs, enhances their independence and competence, and leads to positive results with respect to audit quality. However, previous studies show that compared with smaller accounting firms, larger accounting firms have higher levels of independence and audit quality; this is ascribed to heightened risks of litigation and reputational harm (Francis & Krishnan, 1999;Ireland & Lennox, 2002;Li & Luo, 2011;Qi et al., 2004;Francis & Yu, 2009). What this suggests is that positive impacts of accounting firms' party organisation construction on auditor independence and competence might be less pronounced in larger accounting firms. Conversely and in general, auditor independence and competence are lower in smaller accounting firms. In this case, one might expect the positive effect of party organisation construction on auditor independence and professional competence and, in turn, on audit quality to be greater and more pronounced in smaller firms. On the other hand, audit practice also reveals that many clients eschew highquality external audit services, preferring instead smaller accounting firms of lesser independence and competence. Given this, quality improvements in smaller accounting firms might reasonably be expected to result in loss of clientele which would amount to weakening the positive influence of party member auditors in small firms. Thus, the issue that whether and how the audit firms' size affects the influence of audit firms' party construction on audit quality remains to be verified. To address it, we divide the sample into two groups (namely, Big4 and Non-Big4), and re-estimate the regression model and results are shown in Table 8.  Panel A of Table 8 presents the results of the impact of Big4 audit firms' party organisation construction on audit quality. In columns (1), (3), and (5) of Panel A, the coefficients on PARTY are insignificant. In columns (2), (4), and (6) of Panel A, the coefficients on POST1× PARTY and POST2× PARTY are all insignificant, suggesting that the positive impact of accounting firms' party organisation construction on partymember audit quality is insignificant within the Big4 group. Panel B of Table 8 presents the results of the impact of Non-Big4 audit firms' party organisation construction on audit quality. In columns (1), (3), and (5) of Panel B, the coefficients on PARTY are insignificant. In columns (2), (4), and (6) of Panel B, the coefficients on POST1× PARTY are all insignificant, but the coefficients on POST2× PARTY coefficients are significant. These results indicate that the positive impact of audit firms' party organisation construction on party-member audit quality is significant in the Non-Big4 group. Overall, our findings show that the positive impact of accounting firms' party organisation construction on the audit quality of party member-auditors is more pronounced in smaller accounting firms.

The effect of the proportion of party-member auditors in accounting firms
As stated above, party organisation construction within accounting firms may improve audit quality by strengthening political beliefs amongst party members. If so, then the higher the ratio of party members in accounting firms, the greater the audit quality under conditions of party organisation construction. We conduct an extended test to further query this idea. Specifically, using CICPA data from 2010 to 2012, we take the A-share listed companies from that timespan as objects to examine the relationship between the proportion of party members in accounting firms and audit quality. The regression model is as follow [2] Where Audit Quality is proxied by three different measures, including accrual quality (AbsDA), audit reporting aggressiveness(ARAgg) and accounting conservatism (C_Score).
Psf_ratiao and Pps_ratiao are explanatory variables. Psf_ratiao equals the proportion of the number of party members divided by the total number of employees, and Pps_ratiao equals the proportion of the number of party-member partners (shareholders) divided by the total number of partners (shareholders). Controls variables are similar to those in model (1) and (2). Results are displayed in Table 9. In column (1), the coefficient on Psf_ratiao is −0.042 and significant at 10% level (t-value = −1.702), indicating that the higher the proportion of party-member employees in accounting firms, the higher level of clients' accrual quality. In column (3), the coefficient of Psf_ratiao is −0.070 and significant at 10% level (t-value = −1.704), suggesting there is a significant negative relationship between the proportion of party-member employees in audit firms and aggressive audit reports. In column (5), the coefficient of Psf_ratiao is 0.005 and insignificant (t-value = 0.351), suggesting there is no significantly relationship between the proportion of party-member employees in audit firms and clients' accounting conservatism. These findings confirm that the greater the proportion of party members in accounting firms, the higher the audit quality. In column (2), the coefficient of Pps_ratiao is −0.012 and significant at 5% level (t-value = −2.082), indicating that the more party-member partners (shareholders) per capita in accounting firms, the higher the clients' accrual quality. In column (4), the coefficient of Pps_ratiao is −0.043 and significant at 10% level (t-value = −1.683), suggesting that the more party-member partners (shareholders) per capita in accounting firms, the lower level of audit reports aggressive. In column (6), the Pps_ratiao coefficient is −0.003 and insignificant (t-value = −0.831). Those results show that the higher the ratio of party-member partners (shareholders) in accounting firms, the higher the audit quality becomes. Overall, our findings suggest that in the context of party organisation construction, as the proportion of party members in audit firm increases, so does the audit quality. This table reports results of the relationship between the ratio of party members of accounting firms and audit quality. t-statistics shown in parentheses. *, **, and *** indicate significance at the 10%, 5%, and 1% levels, respectively.

The effect of firm property rights
Audit quality is understood to be the joint probability that an auditor discovers and then reports any irregularities in a client's accounting information system (DeAngelo, 1981). It is thus doubly affected by an auditor's professional competence, and his/her independence (DeFond & Zhang, 2014; Watts & Zimmerman, 1983). In turn, professional competence is determined by auditor characteristics (DeFond & Francis, 2005;DeFond & Zhang, 2014;Gul et al., 2013;Li et al., 2017), while independence is more a function of the institutional environment Wang et al., 2008). This means that both macro and micro systems affect auditor independence, and hence audit quality. The nature of firm property rights has been the most basic micro-institutional arrangement over the course of China's transition period, which has led to fundamental differences in the relationship between the government on one hand, and state-owned vs. non-state-owned enterprises on the other (Luo & Tang, 2009). Due to the unique property rights' characteristics and institutional arrangements of state-owned enterprises, their motivations for financial reporting and demands for high-quality audit services differ from those of non-state-owned enterprises (Chen, 2014). Previous studies show that state-owned companies in the area with lower market-based degree are more likely to hire smaller local firms. Further, in this domain the government is also more interventionist with respect to auditors, thereby impinging more upon the auditor independence (Wang et al., 2008). If so, do firm property rights affect the positive impact of accounting firm party organisation construction on the audit quality of partymember auditors?
To examine impact variation between state-owned and non-state-owned enterprises, we divide the sample into two groups, namely, state-owned enterprises (SOE) and non-state-owned enterprises(Non-SOE). Regression results are shown in Table 10. Panel A of Table 10 reports the results of the impact of audit firms' party organisation construction on audit quality in SOE. In columns (1), (3) and (5), the coefficients on PARTY are insignificant. In columns (2), (4), and (6), the coefficients on POST1× PARTY are insignificant, but the coefficients on POST2× PARTY are all significant, which are similar to previous results. Those findings suggest that the positive impact of audit firms' party organisation construction on party-member audit quality is significant in the SOE. Panel B of Table 10 reports the results of the impact of audit firms' party organisation construction on audit quality in Non-SOE. In columns (1), (3) and (5), the coefficients on PARTY are insignificant. In columns (2), (4), and (6), the coefficients on POST1× PARTY and POST2× PARTY are all insignificant, suggesting that the positive impact of party organisation construction on party-member audit quality is insignificant in the Non-SOE. Our findings indicate that the positive impact of audit firms' party organisation construction on partymember auditors' audit quality is more pronounced in SOE.

Conclusion
Taking publicly traded, Chinese A-Shared market companies from the period 2004 to 2016 as our research sample (and manually collecting information on auditor backgrounds from the CICPA website), we examine the impact of accounting firms' party organisation construction on audit quality. We conclude that although there is no significant difference in audit quality between party and non-party member auditors in general, party-member audit quality improves significantly after audit firms strengthen party organisation construction. Our conclusion remains valid after a series of robust tests. Further research indicates that the positive impact on audit quality of auditors' political beliefs does not have a contagious effect, given the presence of party organisation construction within the firm. That is, compared with non-party auditors lacking experience of cooperation with party auditors over the previous year, non-party auditors with cooperation experience did not demonstrate improved audit quality. As for the heterogeneity of accounting firms, we find that compared with Big4 accounting firms, the positive impact of party organisation construction on audit quality is more pronounced in Non-Big4 firms. Moreover, we find that given party organisation construction within a audit firm, the higher the proportion of both party-member employees and partners (shareholders), the higher the audit quality can be expected to be. Finally, considering differences in clients' property rights, we find that accounting firms' party organisation construction plays a more significant quality-improvement role in state-owned enterprises.
Our paper has major practical and policy significance. We show that party organisation construction within accounting firms can strengthen party-member political beliefs, which in turn leads to improvements in audit quality. This finding not only supports the intention of regulators to promote all-round development of accounting firm business and quality through party organisation construction, but also illuminates a path towards further strengthening accounting firm party organisation construction.