2. Auditor-CEO Surname Sharing and Financial Misstatement

In China, at least two auditors (three auditors in some cases) are required to sign their full names in the client’s audited financial statements. One of them is known as the engagement (field) auditor, and the other is the review auditor. The review auditor must be an audit partner, and the engagement auditor does not have to be an audit partner.

Auditor-client relation refers to the cases in which former audit partners of an accounting firm serve as a client firm’s senior directors or top managers. The data on auditor-client relation can be obtained by checking the resumes of senior directors or top managers, which are provided in CSMAR (China Stock Market and Accounting Research) database.

In most Asian countries such as China, Korea, Japan, and Vietnam and several European and Latin American countries, one’s surname is placed before his/her given name. However, in most Western countries, one’s surname is accustomed to being placed after his/her given name (Doll 1992). It is common that a person has the sole surname, but a Spaniard or Portuguese may own two or more surnames (in rare cases, people do not have to have a surname).

As a result, the vast majority of Chinese people in contemporary society—both men and women, before marriage and after marriage—do not change their surnames during their whole lives (Bai and Kung 2014).

In a guanxi-based society, individual behavior is shaped by interpersonal relationship, and thus some transactions depend on personal agreements and private information, rather than formal contracts (Li 2003). As a comparison, a rule-based society emphasizes publicly verifiable information and the contractual spirit (Li et al. 2004).

As a comparison, in most countries, researchers cannot obtain individual auditor-specific data such as names and other demographic features of the signing auditors or audit partners. To the best of my knowledge, data on individual auditors can be obtained only in few countries or regions such as mainland China, Taiwan (China), and Australia, etc.

In European countries, surnames are associated with a medieval naming movement, which aimed at coping with some situations in which several people had an identical name.

Extant studies have explored the effect of different cultural dimensions on economic development (Barro and McCleary 2003), corporate governance (Du 2015), financial decisions (Hilary and Hui 2009), earnings quality (Du et al. 2015; McGuire et al. 2012), and audit behavior (Gul and Ng 2018; Omer et al. 2018).

“Categorization” implies that people classify themselves into a group based on some common features and “identification” means that people consider themselves to have common characteristics of a given group (Tajfel 1970).

According to DeAngelo (1981a), audit quality is defined as the joint probability of: (1) auditors’ capability of discovering a breach in clients’ financial reports; and (2) their willingness of disclosing this breach.

Severe regulatory penalty and/or reputation losses may hinder audit failures and reduce the risk of financial misstatement in well-developed and highly concentrated audit markets. However, the Chinese audit market is less concentrated and highly competitive (Choi and Wong 2007; Chu et al. 2011; Du 2019), and thus relatively weaker reputation mechanism and litigation effect are less likely to play their restraining roles in financial misstatement. Untabulated results reveal that a firm’s financial misstatement and financial irregularities: (1) do not cause the audit firm (the signing auditors) to undergo the penalty from regulatory bodies; and (2) do not result in the decrease in the market share of the audit firm and the signing auditors.

Under the principal-agent framework (Jensen and Meckling 1976), the CEO belongs to the agent. Independent audit as an external monitoring mechanism is expected to restrain the CEO from pursuing personal benefit at the expense of shareholders (Choi and Wong 2007). However, auditor-CEO social ties (e.g., school ties, teacher-student relation, etc.) may weaken the monitoring role of external audit (Guan et al. 2016; Du 2019). Therefore, this chapter focuses on auditor-CEO surname sharing (ACSS) to address the concern about whether ACSS impairs auditor independence and increases the risk of financial misstatement. In addition, the audit committee as an internal governance mechanism—which is in charge of monitoring the client firm’s internal audit procedures (system) and accounting policy choices—can also play a monitoring role in mitigating the CEO’s unethical behaviors and moral risk. Thus, on the one hand, the audit committee and external audit interact and monitor the CEO. On the other hand, surname sharing between auditor and audit committee members may strengthen or weaken the monitoring effects (Cohen et al. 2010b; DeZoort et al. 2006; He et al. 2017; Rose 2007; Uzzi 1996). In other words, the impact of surname sharing between the signing auditors and audit committee members on auditor independence (audit quality) is mixed at best, and thus it is an empirical issue in essence. Above argument can borrow the support from He et al. (2017). Specifically, focusing on alma mater connections, professor-student bonding, and employment affiliation, He et al. (2017) also argue that two competitive channels determine the effects of social ties between auditors and audit committee members on audit quality. As a result, for the sake of research focus, this chapter focuses on auditor-CEO surname sharing and only includes surname sharing between auditor and audit committee members as a control variable. For similar reasons, surname sharing between the CEO and audit committee members is also included as another control variable.

Referring to Wang et al. (2008), audit firm fixed effects have been taken into account the merge of audit firms.

Analogously, this chapter differentiates the shared rare surname from the shared common surname between the signing auditors and audit committee members, and further two additional variables of AUD_ACM_RARE and AUD_ACM_COM are constructed (see Appendix 2.1 for details)

Please note that results using 70% and 80% as thresholds of the cumulative percentage are qualitatively similar to those using 75% as the threshold. In addition, please note that there are very few firm-year observations for ACSS_RARE if this chapter adopt stricter thresholds, which results in the impossibility of testing Hypothesis 2.

Please refer to Article #54, Section #6, and Chapter #3 of “Notice on issuing the guidelines for the governance standards of Chinese listed firms” (the China Securities Regulatory Commission (CSRC), 2012-01-07; see the official website of CSRC: http://​www.​csrc.​gov.​cn/​pub/​newsite/​ssb/​ssflfg/​bmgzjwj/​ssgszl/​200911/​t20091110_​167722.​html).

The CEO refers to the general manager, rather than the chairman. Firm-years with CEO-chairman duality are deleted for the following reasons: (1) In general, the CEO, the chairman, and the signing auditors (external audit) represent the agency, internal monitor, and external monitor, respectively. (2) Extant studies (Choi and Wong 2007) suggest that external audit can serve as an external monitoring and governance mechanism to constrain the CEO. However, it is validated in previous studies (Du 2019) that social ties between the CEO and the signing auditors may weaken the role of external audit in monitoring the CEO. As a response, extending the existing literature, the aim of this chapter is to investigate whether auditor-CEO surname sharing as a specific social tie may impair auditor independence and lead to financial misreporting (misstatement). (3) In general and theoretically, the chairman can also serve as an internal governance mechanism to monitor the CEO. Therefore, the signing auditors (external audit) as an external governance mechanism and the chairman as an internal governance mechanism can interactively monitor the CEO. (4) CEO-chairman duality means that the CEO can conduct self-monitoring. In this regard, with regard to the restraining effect on the CEO, if firm-year observations with CEO-chairman duality are included, then it may be difficult to distinguish the governance role of the chairman from the monitoring effect of external audit. Based on above reasons, in main tests, this chapter deletes firm-year observations with CEO-chairman duality. Nevertheless, in robustness checks of Table 2.8, this chapter includes firm-years with CEO-chairman duality and finds qualitatively similar results.

Results are qualitatively similar when the top (bottom) 1% of each continuous variable are deleted.

First, this identifies the period in which a financial misstatement happens. Second, I classify and identify the specific type for each financial misstatement, including financial misstatements related with balance sheet, financial misstatements associated with income statement, reclassification-related misstatements, financial misstatements due to new accounting standard adoption, and etc. Third, for each misstatement event, the amount of financial misstatement is calculated. Finally, only financial misstatements associated with income statement are included in this chapter to define the variable of MIS_DUM and MIS_MAG. Using the same procedures, this chapter can obtain data on financial overstatement (OVER_DUM and OVER_MAG). Untabulated results show that the number of events about financial misstatements affecting balance sheet, financial misstatements associated with income statement, reclassification-related misstatements, and financial misstatements due to new accounting standard adoption is 1,914, 1,480, 244, and 103, respectively. Specifically, there are 1,480 restatement events associated with income statement during my sample period (2002–2012), covering 1,037 unique firms. As a result, the mean value of MIS_DUM is 0.079 (1,037/13,128). Based on the aforementioned discussions, the magnitude of financial misstatement (MIS_MAG) is measured as the amount of financial misstatement (due to events associated with income statement) for a firm’s financial statements in year t scaled by total assets (× 100). Similarly, the magnitude of financial overstatement (OVER_MAG) is measured as the amount of overstatement (due to events associated with income statement) for a firm’s financial statements in year t scaled by total assets (×100).

The asymmetric effects can be explained as below: First, the CEO and the audit committee play different roles in contemporary enterprises. Specifically, the CEO serves as the role of agent, but audit committee serves as a governance mechanism to monitor the client firm’s accounting policy choices and internal audit procedures, and thus audit committee can monitor the CEO to some extent. Second, given the external monitoring role of independent auditors (Choi and Wong 2007), auditor-CEO surname sharing as a specific social tie may impair auditor independence, attenuate the monitoring effect of external audit, and thus result in higher risk of financial misstatement. Finally, with regard to monitor the CEO (as well as internal control system), external audit and audit committee may interact and strengthen the monitoring effect. In this regard, a branch of extant studies in the fields of communication and administrative science have argued and found that social ties may facilitate information transfer and flow through the reinforced trust between some established social connections (Gibbons 2004; Sias and Cahill 1998). Extending above findings, surname sharing between the signing auditors and audit committee members is likely to lead to better information communication about internal control environment (weaknesses, strengths), the existing accounting system, tones at the top, and managerial style (Rennie et al. 2010), which is beneficial to the auditing efficiency and audit quality, and eventually reduces the risk of financial misstatement. Using surname sharing between the signing auditors and audit committee members as a control variable, this chapter provides some support to above arguments—surname sharing between the signing auditors and audit committee members can mitigate the risk of financial misstatement.

Specific to surname sharing between the signing auditors and audit committee members (AUD_ACM), its influence on the risk of financial misstatement is mixed, which is an empirical question in essence. On the one hand, extant studies (Gibbons 2004; Sias and Cahill 1998) suggest that social ties can facilitate people to communicate efficiently and even allow them to discuss sensitive topics that mean unthinkable and unspeakable for people without social ties. In addition, Lewicki and Bunker (1996, 114–119) argue that, compared with calculus-based trust and knowledge-based trust, identification-based trust on the basis of social ties is the highest level of interpersonal trust, in which individual identifies himself (herself) with others’ preferences, desires and intentions. Even, in some cases, one’s interests can be respected and protected in interpersonal relationship on the basis of social ties, and thus it seems that some monitoring mechanisms are unnecessary or they cannot work well (Lewicki and Bunker 1996, 114–119). Therefore, AUD_ACM can result in less information exchange barrier (even barrier free) between them (He et al. 2017). In addition, AUD_ACM can back the signing auditors up to resist the pressure from the CEO and top managers (Cohen et al. 2010b; DeZoort et al. 2006). In this regard, AUD_ACM may mitigate the potential audit error and reduce the risk of financial misstatement. On the other hand, due to the worries about the impairment of the established social ties, the signing auditors sharing the same surname with audit committee members may not consciously or unconsciously challenge the potential deficiencies in a client firm’s internal control system and accounting policy choices that the audit committee should be responsible for (He et al. 2017). As a result, surname sharing with audit committee members may hinder the signing auditors from keeping due diligence and weaken external monitoring (Du 2019; He et al. 2017). Also, surname sharing may induce the signing auditors’ unwarranted trust in audit committee members and the supportive materials provided by audit committee members, which leads to foreseeable biases into professional judgment and audit decisions (He et al. 2017; Rose 2007; Uzzi 1996). Thus, AUD_ACM may impair audit quality and increase the likelihood of financial misstatement. To sum up, since different types of social ties may have asymmetric effects on accounting and auditing behaviors, the influence of AUD_ACM on financial misstatement belongs to an empirical issue. Nevertheless, this chapter includes AUD_ACM as a control variable, and the findings show that AUD_ACM reduces the likelihood of financial misstatement.

See the official website: http://​www.​gov.​cn/​zwgk/​2010-07/​23/​content_​1662348.​htm.

Results using a sample period of 2007–2013 excluding 2010 (i.e., three years before and after the event) (Lennox 2016) are qualitatively similar to those using a sample period of 2008–2012 excluding the event year of 2010.

Liu et al. (2012) identify the measure of isonymy, proxied by the surname proximity across different provinces in China. Directors on corporate boards are likely to be chosen locally (Knyazeva et al. 2009, 2013), so surname proximity can embody the isonymy in an area where a client firm is located. The same can be applied to audit firms.

The matching range of ±0.01 is chosen in order to retain as many treated firm-year observations as possible, or else it will lead to lose too many treated firms as unmatchable. As a matter of fact, ±0.01 is a generally accepted standard for the propensity score matching process (Dehejia and Wahba 2002).

If the caliper is tightened (e.g., ±0.005), results remain qualitatively similar to those using a range of ±0.01, although the number of the treated firms varies accordingly.

According to the Chinese Independent Auditing Standards (CIAS), the types of audit opinions include: (1) unqualified opinions; (2) unqualified opinions with explanatory notes; (3) qualified opinions; (4) disclaimer opinions; and (5) adverse opinions. Referring to extant studies (Chen et al. 2000), types (2)–(5) are identified as modified audit opinions (please refer to CIAS No #1502: “Non-standard Auditing Report”).

To examine whether auditor-CEO surname sharing (ACSS) is associated with auditor-CEO hometown relationship, this chapter conducts Pearson correlation analyses between ACSS and three variables of hometown relationship. Untabulated results show that, on the whole, auditor-CEO hometown relationship is insignificantly related to the proxies for auditor-CEO surname sharing. As a result, auditor-CEO hometown relationship is not necessarily completely equivalent to auditor-CEO surname sharing. This finding can borrow important support from Du and Yuan (1993), who argue that several population migrations in the Chinese history result in the decentralization of people with the same surname in different provinces (prefectures, counties) in China.

To collect the data on hometown relationship, researchers have to obtain the original information from personal identification cards of the CEO and the signing auditors. However, as Du (2019) notes, only a small proportion of personal identification cards of the CEO and the signing auditors can be obtained through data mining. Therefore, similarly, a large number of missing data on auditor-CEO surname sharing (ACSS) (non-tabulated for brevity) hinder this chapter from isolating the impact of hometown relationship from beginning to end. Specifically, as shown in Panel A of Table 2.13, only 2,287 firm-year observations (521 unique firms) are available. As a result, this chapter only includes auditor-CEO hometown relationship to conduct additional tests, rather than main tests.

There are very few firm-years with auditor-CEO rare surname sharing.