Connected at the Top: CEO–Executive Social Ties and Audit Pricing

Abstract

This study examines whether social ties within top management teams are associated with audit pricing. Drawing on research on organizational social networks and auditor risk assessment, I investigate whether connections between chief executive officers and other top executives are reflected in auditors’ pricing. Using 22,220 S&P 1500 firm-year observations from 2002 to 2014, I find that CEO–executive connections are negatively associated with audit fees. The findings are consistent with connected executive teams being associated with lower audit risk or client business risk, or with auditors treating such teams as lower-risk clients. The negative association between CEO–executive connections and audit fees is weaker among firms facing higher litigation risk, suggesting that auditors’ pricing responses to organizational social structure are constrained when professional and legal risks are salient. In contrast, the negative association is stronger when auditors have longer tenure and when auditors are industry specialists, consistent with auditors being more likely to incorporate CEO–executive connections into audit pricing when they have greater client-specific or industry-specific knowledge. Additional analysis indicates that the negative association is more consistently observed for advice-based ties than for friendship-based ties, consistent with the view that work-related social ties facilitate information sharing, communication, and coordination within the top management team. Overall, the study contributes to the audit pricing and auditor risk-assessment literature by providing archival evidence that informal organizational ties within the client’s top management team are associated with audit pricing outcomes.

1. Introduction

Organizations are not only formal hierarchies; they are also social systems. Social ties shape information transfer, trust, influence, monitoring, and coordination within organizations (Granovetter, 1973; Uzzi, 1996). Upper echelons theory suggests that organizational outcomes are shaped by the characteristics, experiences, and interactions of top executives (Hambrick & Mason, 1984; Hambrick, 2007). Within top management teams, executives may be connected through prior employment, education, board service, clubs, charitable organizations, and other shared affiliations. These ties represent an informal social structure that can affect how executives communicate and how external observers interpret the organization (Westphal, 1999; Hwang & Kim, 2009; Balsam & Kwack, 2022).

This study examines whether social ties between the chief executive officer (CEO) and other top executives are associated with audit pricing. Auditing is a useful setting for studying how client-level organizational characteristics are reflected in external audit outcomes because auditors evaluate client risk using both formal accounting evidence and qualitative information about the client, management, governance, and the control environment (Bell et al., 2001; Bedard & Johnstone, 2006; Hogan & Wilkins, 2008; DeFond & Zhang, 2014). Recent studies examining financial reporting decisions have begun to consider the role of executives beyond CEOs and CFOs (Kwak et al., 2012; Ge et al., 2011). In addition, Lauck et al. (2020) document that executive fixed effects explain a meaningful portion of variation in audit fees, representing between 20 and 39 percent of the overall variation in unexplained audit fees. This evidence highlights the importance of executive characteristics in audit engagements and suggests the need for further research on whether characteristics of the broader top management team, rather than only those of the CEO or CFO, are reflected in audit pricing. However, relatively little is known about whether the social structure of the broader top management team is associated with auditors’ risk assessments.

Audit standards provide an institutional basis for examining whether informal features of top management are relevant to auditors’ risk assessments. PCAOB AS 2201 requires auditors, when evaluating the control environment, to assess whether management’s philosophy and operating style promote effective internal control and whether integrity and ethical values, particularly those of top management, are developed and understood (PCAOB, 2007, para. 25). The standard also recognizes that “documentary evidence of the operation of some controls, such as management’s philosophy and operating style, might not exist” (PCAOB, 2007, para. 51). These provisions suggest that auditors’ assessments may extend beyond formal accounting systems to less visible managerial attributes and organizational characteristics. In this study, CEO–executive social ties are examined as one such informal feature of the top management team.

This study addresses this issue by examining whether CEO–executive social ties are associated with audit pricing. Auditors have incentives to consider client-level organizational features for several reasons. First, auditors have reputations to protect and are subject to litigation risk. Reputation concerns provide auditors with incentives to conduct high-quality audits because reputation affects their ability to attract and retain clients (DeFond & Zhang, 2014). Litigation risk also motivates auditors to increase audit effort because of the potential financial and reputational costs associated with audit failure (Simunic & Stein, 1996). Accordingly, auditors have incentives to identify inherent and control risks that may affect audit effort and pricing. Second, auditors are relatively well positioned to obtain information about the backgrounds and interactions of top executives. Although shareholders and investors may obtain executive background information from 10-K filings or proxy statements, these disclosures are limited to information the firm chooses to report. Auditors, by contrast, have access to additional information channels, including interactions with the audit committee1. Through such interactions, auditors may obtain information about executive backgrounds, management style, and governance concerns that is less visible to outside investors. Therefore, CEO–executive social ties may represent an informal organizational feature that is relevant to auditors’ client risk assessments and reflected in audit pricing.

Prior research provides competing predictions about how CEO–executive social ties relate to audit risk and client business risk, with some studies pointing to risk-mitigating effects through improved coordination, information sharing, and mutual monitoring (Gaspar & Massa, 2011; Ke et al., 2019; Balsam & Kwack, 2022), and others pointing to risk-increasing effects through weaker independent challenge and greater collusive potential (Khanna et al., 2015; Zhang, 2019). Audit fees provide an observable outcome through which to examine these competing implications, since audit fees are known to reflect client business risk, auditor business risk, and audit effort (Bell et al., 2001; Bedard & Johnstone, 2006). Given these competing views, the direction of the association between CEO–executive social ties and audit fees is unclear ex ante.

I test these competing interpretations using 22,220 S&P 1500 firm-year observations from 2002 to 2014. CEO–executive social ties are measured using BoardEx by identifying whether the CEO and each executive share employment, educational, or other social affiliations. The test variable is the percentage of executives who are socially connected to the CEO. I find a negative association between CEO–executive connections and audit fees. This evidence is consistent with auditors pricing engagements involving more connected top management teams as lower risk. As the archival design does not directly observe auditors’ risk assessments, the evidence is also consistent with connected teams having lower underlying business or financial-reporting risk that auditors price similarly. I explore this distinction further through cross-sectional tests. I find that the negative association between CEO–executive ties and audit fees is weaker for firms in high-litigation industries. This finding is consistent with litigation pressure acting as a boundary condition: when legal exposure is high, auditors have stronger incentives to be cautious in translating potentially favorable informal organizational characteristics into lower audit fees (Krishnan & Krishnan, 1997; Venkataraman et al., 2008; Choi et al., 2009; Abbott et al., 2017). In contrast, the negative association is stronger when auditors have longer tenure and when auditors are industry specialists. These findings are consistent with CEO–executive connections being more strongly reflected in audit pricing when auditors have greater client-specific or industry-specific knowledge. In additional analyses, I distinguish advice-based ties from friendship-based ties following Bruynseels and Cardinaels (2014). The results suggest that the associations are more evident for advice-based ties than for friendship-based ties, consistent with the view that work-related social ties facilitate information flow, communication, and coordination within the top management team. Finally, the results remain robust to a two-stage instrumental variable estimation, a propensity score-matched (PSM) sample, and a propensity score-matched difference-in-differences analysis (PSM-DID) around CEO turnover.

This study makes three contributions. First, this study contributes to the audit pricing and auditor risk-assessment literature by examining whether informal organizational ties within the client’s top management team are associated with audit fees. Prior audit fee research shows that audit fees reflect client risk, auditor effort, auditor business risk, and governance-related monitoring demand (Simunic, 1980; Bell et al., 2001; Carcello et al., 2002; Bedard & Johnstone, 2006; Hogan & Wilkins, 2008). This study extends that literature by identifying CEO–executive connections as a less visible organizational feature associated with audit pricing. The evidence is consistent with the interpretation that CEO–executive connections are reflected in audit pricing outcomes in a manner consistent with auditors’ engagement-risk assessments.

Second, this study contributes to research on social networks and organizational behavior. Prior research shows that social ties influence information transfer, trust, coordination, social capital, and monitoring within organizations (Granovetter, 1973; Uzzi, 1996; Nahapiet & Ghoshal, 1998; Borgatti & Foster, 2003; Brass et al., 2004). While prior studies on CEO–executive connections primarily examine internal outcomes such as disclosure, investment, firm value, fraud, and reporting quality (Gaspar & Massa, 2011; Khanna et al., 2015; Ke et al., 2019; Zhang, 2019; Balsam & Kwack, 2022; Kuang et al., 2022), this study shows that such ties are also associated with external audit pricing. This suggests that internal social structures within the top management team may have implications beyond the boundaries of the client organization.

Third, this study contributes to auditing and corporate governance research by identifying a management-level characteristic reflected in audit pricing. Prior auditing research has focused largely on formal governance structures, such as boards and audit committees (Carcello et al., 2002; Abbott et al., 2003; Bruynseels & Cardinaels, 2014; DeFond & Zhang, 2014). This study complements that literature by showing that informal characteristics of the broader executive team are associated with audit fees after controlling for formal governance characteristics. This evidence is consistent with audit standards emphasizing management philosophy, operating style, integrity, ethical values, and the control environment (PCAOB, 2007).

In addition to its contributions to the academic literature, the study has implications for audit firms, audit committees, and regulators. For audit firms, the results suggest that top management team social ties may be useful inputs to audit methodology and staff training related to client risk assessment. Training should emphasize that such ties may have both risk-reducing and risk-increasing implications, and engagement teams should evaluate CEO–executive connections together with other client-specific factors, such as governance quality, litigation exposure, auditor tenure, and industry expertise. For audit committees, the findings suggest that greater transparency about senior executives’ backgrounds and relationships may help auditors and audit committees better understand the client’s internal reporting environment. For regulators, the results are relevant because AS No. 5/AS 2201 recognizes that some controls related to management may not leave formal documentation. The tone set by top executives can have a substantial impact on the firm’s internal control (Ge & McVay, 2005; Schmidt, 2014) and has been linked to some of the financial frauds committed in the last two decades (Schwartz et al., 2005). Thus, CEO–executive connections represent the type of informal organizational feature that may be difficult to observe directly, yet still relevant to auditors’ risk assessments. However, because the evidence is archival and based on the 2002–2014 institutional setting, these implications should be interpreted cautiously and should not be viewed as supporting a specific new disclosure mandate or contemporaneous policy recommendation.

The remainder of the paper is organized as follows. I review the related literature and develop hypotheses in Section 2. Then the data sources and research design are described in Section 3, and the empirical results are presented in Section 4. Section 5 presents the additional analyses and robustness tests. Section 6 discusses the theoretical implications and limitations, and Section 7 concludes.

2. Theoretical Background and Hypothesis Development

This section develops the hypotheses by linking research on organizational social ties to auditors’ assessments of client risk. Social network research in accounting and finance shows that social ties can affect information transfer, monitoring, governance, and financial reporting outcomes (see Bianchi et al., 2023, for a recent review). Recent audit research and regulatory developments also highlight the continuing importance of auditor risk assessment in settings where auditors increasingly use data analytics and technology-assisted analysis to evaluate risk-relevant information (Eilifsen et al., 2020; Koreff, 2022; Brazel et al., 2022; PCAOB, 2024). This study complements this literature by examining whether a less visible organizational feature of the client, the social structure of the top management team, is associated with audit pricing.

CEO–executive connections are informal organizational characteristics that may be relevant to the client’s information environment, management coordination, and reporting risk. Prior research suggests that these connections may have competing implications for audit risk and client business risk. On the one hand, CEO–executive connections may facilitate information sharing, coordination, and mutual monitoring within the top management team, thereby reducing audit or client business risk. Consistent with this view, prior studies find that CEO–executive or CEO–manager connections are associated with more accurate voluntary disclosure, improved resource allocation, and higher firm value. For example, Ke et al. (2019) find that CEO connections to executives improve management forecast accuracy, while Gaspar and Massa (2011) and Balsam and Kwack (2022) document positive associations between CEO–manager or CEO–executive connections and firm value. These findings suggest that CEO–executive connections may be associated with lower audit risk or client business risk.

On the other hand, they may signal reduced independent challenge, greater consensus, or collusive potential, thereby increasing perceived inherent and control risk. Consistent with this view, Khanna et al. (2015) find that greater connections between CEOs and executives/directors increase the likelihood of fraud and decrease the likelihood of detection, while Zhang (2019) shows that top management background homogeneity is associated with more restatements and discretionary accruals. Other studies show that the effect of connections depends on context: CEO–manager connections are beneficial under high information asymmetry but costly under weak governance (Duchin & Sosyura, 2013), and CFO–senior manager connections are associated with higher or lower restatement likelihood depending on operational performance (Kuang et al., 2022). Thus, prior literature supports both risk-mitigating and risk-increasing interpretations of CEO–executive connections.

I examine whether CEO–executive connections are associated with audit fees, an observable outcome of auditor risk assessment and effort decisions. Audit fees reflect the amount of client risk, auditor business risk, and audit effort put in by the audit firm (Bell et al., 2001; Bedard & Johnstone, 2006). When inherent or control risk is higher, auditors increase audit effort and reduce detection risk to maintain the desired level of overall audit risk, resulting in higher audit fees (Hogan & Wilkins, 2008).2 More recently, Ranasinghe et al. (2023) show that the client business risk premium in audit fees is attenuated when clients reduce their business risk, consistent with audit fees reflecting client business risk.

The direction of this association is further complicated by the fact that audit fees also reflect negotiation between auditors and clients, not only auditors’ risk assessments. Connected executives may either conceal misreporting when connections facilitate collusion and reduce independent monitoring, or support greater assurance when connections facilitate open communication and mutual monitoring (Khanna et al., 2015; Zhang, 2019; Westphal, 1999; Ke et al., 2019). Thus, both audit risk considerations and client-side incentives make the direction of the association between CEO–executive connections and audit fees unclear ex ante. Given this ex ante ambiguity, I do not impose a directional prediction and test the association between CEO–executive connections and audit fees as a two-tailed prediction.

H1. 

Connections between the CEO and executives are associated with total audit fees.

To provide additional evidence on the interpretation of the audit fee results, I later conduct cross-sectional analyses based on litigation risk, auditor tenure, and auditor industry specialization. These analyses examine whether the association between CEO–executive connections and audit fees varies with auditors’ incentives and knowledge, but they are not stated as separate formal hypotheses because they are intended to help interpret the main audit fee association.

Litigation risk provides a setting in which auditors have stronger incentives to avoid audit failure and to be cautious in reducing audit fees. Prior research suggests that auditors facing higher litigation exposure exert greater effort and charge higher audit fees because of the potential financial and reputational costs of audit failure (e.g., Simunic & Stein, 1996; Venkataraman et al., 2008; Choi et al., 2009; Abbott et al., 2017). Therefore, if CEO–executive connections are associated with lower audit fees because they capture lower audit risk or client business risk, the negative association should be weaker when litigation risk is high.

Auditor tenure and auditor industry specialization provide settings in which auditors are more likely to possess greater client-specific or industry-specific knowledge. Prior research suggests that longer auditor tenure can increase auditors’ knowledge of the client’s operations, reporting systems, and business environment, although it may also raise independence concerns (e.g., Myers et al., 2003; Ghosh & Moon, 2005; Carey & Simnett, 2006; Lennox et al., 2014). Similarly, prior research shows that industry specialist auditors possess industry-specific expertise that is reflected in audit pricing, audit quality, and financial reporting outcomes (e.g., Balsam et al., 2003; Mayhew & Wilkins, 2003; Francis et al., 2005; Reichelt & Wang, 2010). Therefore, if CEO–executive connections are relevant to audit pricing because they capture information about the client’s organizational environment, the negative association should be stronger when auditor tenure is longer and when the auditor is an industry specialist.

3. Materials and Methods

3.1. Sample and Data

The initial sample is formed based on the intersection of BoardEx, Execucomp, Risk Metrics, Audit Analytics, and Compustat from 2002 to 20143. BoardEx is used to create the connection measure. To measure connections between CEO and executives, I first match the names on Execucomp to that on BoardEx. Then for each CEO and executive pair, I create a binary variable equal to 1 if the CEO and executive are connected, and 0 otherwise. I define a CEO and an executive as having a connection if they are connected through employment (i.e., working as employees or directors at the same time for the same firm other than the current firm), education (i.e., graduated from the same university within two years of each other), or other activities (e.g., affiliation with the same country clubs or charitable or not-for-profit organizations) as identified on BoardEx4. Then, I aggregate this information to the firm-level and define CEO-EXEC CONN as the percentage of executives who have connections with the CEO. After deleting observations with missing variables, the final sample consists of 22,220 firm-year observations.

3.2. Regression Models

To examine how the connections between the CEO and the executives are associated with audit fees, I estimate the following OLS regression which controls for classical audit fee determinants from the literature (e.g., Huang et al., 2009, 2015; Choi et al., 2010):

LN(AUDFEE)_it = β₀ + β₁ CEO-EXEC CONN_it + β₂ CEO-AC CONN_it + β₃ LN(ASSET)_it
       + β₄ BUSSEG_it + β₅ FRGN_it + β₆ INVREC_it + β₇ ROA_it + β₈ LOSS_it + β₉ LEV_it
        + β₁₀ GROWTH_it + β₁₁ ISSUE_it+1 + β₁₂ EXDISC_it + β₁₃ DEC_it + β₁₄ RESTATE_it
+ β₁₅ REPLAG_it + β₁₆ MW_it + β₁₇ LN(AUDTEN)_it + β₁₈ BIG4_it
         + β₁₉ LN(BOARDSIZE)_it + β₂₀ LN(ACSIZE)_it + β₂₁ INDEPDIR_it+ β₂₂ ACFINEXP_it
   + β₂₃ LITIGATION_it + Year Fixed Effects + Industry Fixed Effects + ε

(1)

where subscripts i and t denote firm and fiscal year, respectively.

The dependent variable is LN(AUDFEE), which is defined as the natural logarithm of audit fees, and the main test variable is CEO-EXEC CONN, which is the percentage of executives with connections to the CEO. A positive coefficient on CEO-EXEC CONN, β₁, would be consistent with an audit fee premium reflecting increased audit and client business risk, and with auditors incorporating the potential for reduced independent challenge among connected executives into their risk assessment. A negative coefficient on CEO-EXEC CONN, β₁, would be consistent with a lower level of assessed audit and client business risk, and with auditors incorporating increased information sharing and mutual monitoring among connected executives into their risk assessment. I also use an alternative test variable, MAJ CEO-EXEC CONN, which is a binary variable equal to 1 if a majority of the executives have connections to the CEO, and 0 otherwise.

I first control for the connections between the CEO and the audit committee directors (CEO-AC CONN) because Bruynseels and Cardinaels (2014) document that these connections are negatively associated with audit fees. Next, I incorporate a set of determinants of audit fees following prior research. Specifically, I control for client complexity using firm size [LN(ASSET)], number of business segments (BUSSEG), and foreign operations (FRGN), as operational complexity is associated with higher audit effort and fees (Simunic, 1980; Simon & Francis, 1988). I control for audit risk using inventories and receivables (INVREC), return on assets (ROA), loss (LOSS), and leverage (LEV), because auditors incorporate risk exposure into audit pricing (Simunic, 1980; Simunic & Stein, 1996). I further control for growth and financing activities (GROWTH and ISSUE), other client characteristics (EXDISC, DEC, and RESTATE), auditor characteristics and audit outcomes (REPLAG, MW, LN[AUDTEN], and BIG4), board and audit committee characteristics (LN[BOARDSIZE], LN[ACSIZE], INDEPDIR, and ACFINEXP), and litigation risk (LITIGATION) (Abbott et al., 2003; Carcello et al., 2002; Cho et al., 2021). Finally, I include year and industry fixed effects, with industries defined using two-digit SIC codes.

4. Results

4.1. Descriptive Statistics

Table 1. Descriptive Statistics (N = 22,220).

Variable	Mean	Std. Dev.	Q1	Median	Q3
Dependent Variables
LN(AUDFEE)	13.872	1.205	13.039	13.845	14.649
Connection Variables
CEO-EXEC CONN	0.121	0.232	0	0	0.125
CEO-AC CONN	0.107	0.240	0	0	0
MAJ CEO-EXEC CONN	0.088	0.283	0	0	0
MAJ CEO-AC CONN	0.078	0.269	0	0	0
Control Variables
ACFINEXP	0.438	0.276	0.25	0.333	0.667
BIG4	0.810	0.392	1	1	1
BUSSEG	5.994	5.419	3	3	9
DEC	0.708	0.455	0	1	1
EXDISC	0.243	0.429	0	0	1
FRGN	0.271	0.445	0	0	1
GROWTH	0.101	0.308	−0.032	0.065	0.175
INDEPDIR	0.756	0.131	0.667	0.778	0.857
INVREC	0.289	0.236	0.101	0.232	0.411
ISSUE	0.438	0.496	0	0	1
LEV	0.184	0.196	0.007	0.131	0.292
LITIGATION	0.290	0.454	0	0	1
LN(ACSIZE)	1.567	0.198	1.386	1.609	1.609
LN(ASSET)	6.912	1.988	5.570	6.896	8.224
LN(AUDTEN)	1.933	0.605	1.579	1.946	2.303
LN(BOARDSIZE)	2.247	0.251	2.079	2.197	2.398
LOSS	0.280	0.449	0	0	1
MW	0.053	0.223	0	0	0
REPLAG	62.143	16.147	54	60	72
RESTATE	0.139	0.346	0	0	0
ROA	0.008	0.154	−0.009	0.031	0.078

Note: This table presents the descriptive statistics of the data. The definitions of these variables are in Appendix A.

provides the descriptive statistics. All continuous variables are winsorized at the 1st and 99th percentiles to mitigate the impact of outliers. The mean (median) audit fee before taking its natural logarithm (untabulated) is $2,428,430 ($1,030,000), which is similar to what is reported in studies using S&P 1500 firms as the sample (Ho & Kang, 2013). Due to its skewness, I take its natural logarithm [LN(AUDFEE)]. I find that the mean percentage of executives with connections to the CEO is 12.1%, and the percentage of audit committee directors with connections to the CEO is 10.7%. Also, approximately 8.8% and 7.8% of the sample have a majority of executives with connections to the CEO and a majority of audit committee directors with connections to the CEO, respectively. The descriptives of the control variables are similar to what is documented in the prior literature (e.g., Cho et al., 2021; Carey & Simnett, 2006; Abbott et al., 2003).

Table 2. Pearson Correlations.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)
(1) LN(AUDFEE)	1.000
(2) CEO-EXEC CONN	−0.088 ***	1.000
(3) CEO-AC CONN	−0.048 ***	0.653 ***	1.000
(4) MAJ CEO-EXEC CONN	−0.109 ***	0.854 ***	0.569 ***	1.000
(5) MAJ CEO-AC CONN	−0.083 ***	0.590 ***	0.869 ***	0.531 ***	1.000
(6) LN(ASSET)	0.763 ***	0.126 ***	0.139 ***	0.091 ***	0.094 ***	1.000
(7) BUSSEG	0.291 ***	−0.095 ***	−0.096 ***	−0.090 ***	−0.094 ***	0.191 ***	1.000
(8) FRGN	0.230 ***	−0.109 ***	−0.078 ***	−0.098 ***	−0.082 ***	0.036 ***	0.043 ***	1.000
(9) INVREC	−0.150 ***	0.137 ***	0.126 ***	0.147 ***	0.138 ***	−0.052 ***	−0.097 ***	−0.017 **	1.000
(10) ROA	0.191 ***	0.002	0.005	0.017 ***	0.008	0.314 ***	0.109 ***	0.036 ***	0.140 ***	1.000
(11) LOSS	−0.177 ***	−0.034 ***	−0.037 ***	−0.050 ***	−0.034 ***	−0.328 ***	−0.117 ***	−0.001	−0.138 ***	−0.666 ***	1.000
(12) LEV	0.187 ***	−0.001	0.002	−0.020 ***	−0.014 **	0.249 ***	0.071 ***	−0.066 ***	−0.225 ***	−0.064 ***	0.061 ***
(13) GROWTH	−0.046 ***	−0.018 ***	−0.012 *	−0.021 ***	−0.016 **	−0.037 ***	−0.006	−0.006	0.062 ***	0.051 ***	−0.099 ***
(14) ISSUE	0.057 ***	−0.029 ***	−0.003	−0.031 ***	−0.013 *	0.054 ***	0.022 ***	−0.023 ***	−0.051 ***	−0.122 ***	0.060 ***
(15) EXDISC	0.201 ***	−0.039 ***	−0.039 ***	−0.047 ***	−0.054 ***	0.165 ***	0.149 ***	−0.004	−0.127 ***	0.007	0.012 *
(16) DEC	0.085 ***	0.132 ***	0.115 ***	0.105 ***	0.091 ***	0.139 ***	−0.014 **	−0.046 ***	−0.106 ***	−0.067 ***	0.031 ***
(17) RESTATE	−0.010	−0.033 ***	−0.030 ***	−0.023 ***	−0.026 ***	−0.002	0.046 ***	−0.024 ***	−0.013 *	−0.001	−0.007
(18) REPLAG	−0.225 ***	0.071 ***	0.047 ***	0.056 ***	0.056 ***	−0.384 ***	−0.107 ***	−0.015 **	0.104 ***	−0.216 ***	0.237 ***
(19) MW	0.035 ***	−0.042 ***	−0.045 ***	−0.038 ***	−0.037 ***	−0.105 ***	0.005	0.036 ***	0.020 ***	−0.080 ***	0.109 ***
(20) LN(AUDTEN)	0.214 ***	0.004	−0.005	−0.003	−0.018 ***	0.152 ***	0.025 ***	0.051 ***	−0.028 ***	0.026 ***	−0.031 ***
(21) BIG4	0.487 ***	−0.144 ***	−0.123 ***	−0.135 ***	−0.130 ***	0.416 ***	0.166 ***	0.101 ***	−0.218 ***	0.128 ***	−0.144 ***
(22) LN(BOARDSIZE)	0.489 ***	0.118 ***	0.108 ***	0.090 ***	0.069 ***	0.643 ***	0.134 ***	0.017 **	0.019 ***	0.147 ***	−0.210 ***
(23) LN(ACSIZE)	0.326 ***	0.097 ***	0.081 ***	0.085 ***	0.054 ***	0.421 ***	0.134 ***	0.016 **	0.051 ***	0.125 ***	−0.172 ***
(24) INDEPDIR	0.261 ***	0.015 **	0.042 ***	0.007	0.023 ***	0.223 ***	0.025 ***	0.075 ***	−0.022 ***	0.026 ***	−0.058 ***
(25) ACFINEXP	0.197 ***	−0.041 ***	−0.025 ***	−0.045 ***	−0.041 ***	0.127 ***	0.030 ***	0.040 ***	−0.034 ***	0.041 ***	−0.037 ***
(26) LITIGATION	−0.094 ***	−0.181 ***	−0.143 ***	−0.154 ***	−0.128 ***	−0.243 ***	−0.117 ***	0.051 ***	−0.116 ***	−0.160 ***	0.155 ***
Variables		(12)	(13)	(14)	(15)	(16)	(17)	(18)	(19)	(20)	(21)
(12) LEV		1.000
(13) GROWTH		0.014 **	1.000
(14) ISSUE		0.388 ***	0.104 ***	1.000
(15) EXDISC		0.186 ***	−0.098 ***	0.080 ***	1.000
(16) DEC		0.132 ***	0.044 ***	0.078 ***	0.054 ***	1.000
(17) RESTATE		0.006	0.007	0.012 *	0.017 ***	0.010	1.000
(18) REPLAG		−0.070 ***	−0.012 *	−0.012 *	−0.016 **	0.009	−0.025 ***	1.000
(19) MW		−0.018 ***	0.003	0.009	0.025 ***	−0.029 ***	0.036 ***	0.297 ***	1.000
(20) LN(AUDTEN)		0.017 **	−0.063 ***	−0.010	−0.008	−0.031 ***	−0.078 ***	−0.032 ***	−0.036 ***	1.000
(21) BIG4		0.146 ***	0.012 *	0.045 ***	0.093 ***	0.032 ***	0.037 ***	−0.312 ***	−0.055 ***	0.254 ***	1.000
(22) LN(BOARDSIZE)		0.125 ***	−0.048 ***	0.014 **	0.115 ***	0.083 ***	−0.002	−0.259 ***	−0.081 ***	0.100 ***	0.271 ***
(23) LN(ACSIZE)		0.053 ***	−0.061 ***	−0.005	0.104 ***	0.058 ***	0.000	−0.170 ***	−0.049 ***	0.071 ***	0.143 ***
(24) INDEPDIR		0.004	−0.048 ***	−0.004	0.034 ***	0.068 ***	−0.068 ***	−0.088 ***	−0.037 ***	0.161 ***	0.112 ***
(25) ACFINEXP		0.062 ***	0.001	0.040 ***	0.029 ***	−0.002	−0.048 ***	−0.024 ***	−0.009	0.138 ***	0.088 ***
(26) LITIGATION		−0.188 ***	0.046 ***	−0.060 ***	−0.106 ***	−0.192 ***	−0.009	0.011 *	0.037 ***	−0.007	0.002
Variables							(22)	(23)	(24)	(25)	(26)
(22) LN(BOARDSIZE)							1.000
(23) LN(ACSIZE)							0.523 ***	1.000
(24) INDEPDIR							0.179 ***	0.302 ***	1.000
(25) ACFINEXP							0.040 ***	−0.121 ***	0.113 ***	1.000
(26) LITIGATION							−0.208 ***	−0.192 ***	−0.018 ***	0.025 ***	1.000

*** p < 0.01, ** p < 0.05, * p < 0.1.

reports Pearson correlations among the variables used in the empirical analysis. The correlation between LN(AUDFEE) and CEO-EXEC CONN is negative and significant, providing preliminary evidence consistent with H1. The correlations between LN(AUDFEE) and other connection variables are also negative and significant. The correlation matrix does not suggest that multicollinearity is likely to drive the main results.

4.2. Regression Results

Table 3. Connections between the CEO and the Executives and Audit Fees.

Dependent Variable	LN(AUDFEE)
	Column (1)	Column (2)
CEO-EXEC CONN	−0.154 ***
	(−3.50)
MAJ CEO-EXEC CONN		−0.127 ***
		(−4.60)
CEO-AC CONN	−0.010
	(−0.27)
MAJ CEO-AC CONN		−0.018
		(−0.67)
LN(ASSET)	0.503 ***	0.503 ***
	(64.65)	(64.71)
BUSSEG	0.008 ***	0.008 ***
	(4.73)	(4.73)
FRGN	0.205 ***	0.205 ***
	(11.34)	(11.38)
INVREC	0.217 ***	0.218 ***
	(4.12)	(4.13)
ROA	−0.215 ***	−0.212 ***
	(−3.93)	(−3.89)
LOSS	0.116 ***	0.115 ***
	(7.84)	(7.75)
LEV	−0.097 *	−0.097 *
	(−1.76)	(−1.77)
GROWTH	−0.088 ***	−0.089 ***
	(−6.03)	(−6.13)
ISSUE	0.027 **	0.028 **
	(2.51)	(2.55)
EXDISC	0.161 ***	0.161 ***
	(11.07)	(11.02)
DEC	0.093 ***	0.093 ***
	(4.61)	(4.60)
RESTATE	0.049 ***	0.049 ***
	(4.07)	(4.07)
REPLAG	0.005 ***	0.005 ***
	(11.31)	(11.30)
MW	0.364 ***	0.364 ***
	(15.66)	(15.69)
LN(AUDTEN)	−0.014	−0.0130
	(−1.25)	(−1.16)
BIG4	0.354 ***	0.353 ***
	(15.15)	(15.11)
LN(BOARDSIZE)	0.129 ***	0.126 ***
	(2.94)	(2.87)
LN(ACSIZE)	0.086 **	0.087 **
	(2.18)	(2.21)
INDEPDIR	0.315 ***	0.317 ***
	(4.78)	(4.82)
ACFINEXP	0.079 ***	0.079 ***
	(2.86)	(2.89)
LITIGATION	−0.016	−0.016
	(−0.48)	(−0.48)
INTERCEPT	8.223 ***	8.211 ***
	(27.15)	(27.44)
YEAR FE	YES	YES
INDUSTRY FE	YES	YES
Observation	22,220	22,220
Adj. R2	0.817	0.817

Note: This table shows the OLS regression results on the association between the CEO–executive connections and audit fees. The definitions of the variables are in Appendix A. t-statistics in parentheses are based on clustered standard errors at the firm level. ***, **, and * represent significance levels of 1 percent, 5 percent, and 10 percent, respectively.

shows the results of testing whether CEO–executive connections are associated with audit fees (H1). Column (1) reports the result of estimating Model (1), and Column (2) reports the result of estimating Model (1) after replacing CEO-EXEC CONN with MAJ CEO-EXEC CONN. I find the coefficient on CEO-EXEC CONN in Column (1) significantly negative (−0.154, t-stat = −3.50) and the coefficient on MAJ CEO-EXEC CONN in Column (2) significantly negative (−0.127, t-stat = −4.60)5. These results suggest that CEO–executive connections are negatively associated with audit fees, consistent with auditors interpreting CEO–executive connections as an organizational characteristic associated with lower audit risk or client business risk. The coefficients are also economically meaningful. A one standard deviation increase in CEO-EXEC CONN is associated with approximately 3.5 percent lower audit fees, and the coefficient on MAJ CEO-EXEC CONN indicates that firms in which a majority of executives have connections to the CEO have audit fees that are approximately 11.9 percent lower than those without such majority CEO–executive connections. These magnitudes suggest that the association between CEO–executive connections and audit fees is not only statistically significant but also economically meaningful.

Turning to the control variables, the results are generally consistent with prior audit fee research. Firms with greater operational complexity, as reflected in larger size, more business segments, and foreign operations, are associated with higher audit fees (Simunic, 1980; Simon & Francis, 1988). Audit fees are also higher for firms with greater audit risk, including firms with higher inventories and receivables, lower profitability, losses, restatements, material weaknesses, and longer reporting lags (Simunic, 1980; Hogan & Wilkins, 2008). In addition, firms issuing securities, reporting extraordinary or discontinued items, having December fiscal year-ends, and being audited by Big 4 auditors are associated with higher audit fees (Francis, 1984; Choi et al., 2010). Board size, audit committee size, board independence, and audit committee financial expertise are also positively associated with audit fees, consistent with greater governance-related monitoring demand (Carcello et al., 2002; Abbott et al., 2003). Overall, the control-variable results are broadly consistent with audit fees reflecting client complexity, audit risk, auditor characteristics, and governance-related demand for audit effort.

4.3. Cross-Sectional Tests

Table 4. Cross-Sectional Tests.

Panel A. Litigation Risk
Dependent Variable	LN(AUDFEE)
	Column (1)	Column (2)	Column (3)	Column (4)
	LITIGATION = 1	LITIGATION = 0	LITIGATION = 1	LITIGATION = 0
CEO-EXEC CONN	−0.051	−0.171 ***
	(−0.63)	(−3.46)
MAJ CEO-EXEC CONN			−0.087	−0.129 ***
			(−1.50)	(−4.33)
Controls	YES	YES	YES	YES
YEAR FE	YES	YES	YES	YES
INDUSTRY FE	YES	YES	YES	YES
Observation	6450	15,770	6450	15,770
Adj. R2	0.805	0.823	0.805	0.823
Test difference in coefficients on CEO-EXEC CONN/ MAJ CEO-EXEC CONN between LITIGATION = 1 and LITIGATION = 0 firms
p-value	0.010 **		0.080 *
Panel B. Auditor Tenure
Dependent Variable	LN(AUDFEE)
	Column (1)	Column (2)	Column (3)	Column (4)
	LN(AUDTEN) >median	LN(AUDTEN) ≤median	LN(AUDTEN) >median	LN(AUDTEN) ≤median
CEO-EXEC CONN	−0.193 ***	−0.133 **
	(−3.68)	(−2.37)
MAJ CEO-EXEC CONN			−0.154 ***	−0.111 ***
			(−4.40)	(−3.21)
Controls	YES	YES	YES	YES
YEAR FE	YES	YES	YES	YES
INDUSTRY FE	YES	YES	YES	YES
Observation	10,824	11,396	10,824	11,396
Adj. R2	0.825	0.802	0.825	0.802
Test difference in coefficients on CEO-EXEC CONN/ MAJ CEO-EXEC CONN between LITIGATION = 1 and LITIGATION = 0 firms
p-value	0.060 *		0.050 *
Panel C. Industry Expert Auditors
Dependent Variable	LN(AUDFEE)
	Column (1)	Column (2)	Column (3)	Column (4)
	IND_EXPERT = 1	IND_EXPERT = 0	IND_EXPERT = 1	IND_EXPERT = 0
CEO-EXEC CONN	−0.327 ***	−0.137 ***
	(−3.39)	(−2.91)
MAJ CEO-EXEC CONN			−0.218 ***	−0.114 ***
			(−3.64)	(−3.76)
Controls	YES	YES	YES	YES
YEAR FE	YES	YES	YES	YES
INDUSTRY FE	YES	YES	YES	YES
Observation	3313	18,529	3313	18,529
Adj. R2	0.825	0.802	0.825	0.802
Test difference in coefficients on CEO-EXEC CONN/ MAJ CEO-EXEC CONN between LITIGATION = 1 and LITIGATION = 0 firms
p-value	0.000 ***		0.000 ***

Note: This table shows how litigation risk, auditor tenure, and industry expert auditors moderates the association between CEO–executive connections and audit fees in Panels A, B, and C, respectively. The definitions of the variables are in Appendix A. t-statistics in parentheses are based on clustered standard errors at the firm level. ***, **, and * represent significance levels of 1 percent, 5 percent, and 10 percent, respectively.

reports the results of cross-sectional tests based on litigation risk, auditor tenure, and auditor industry specialization. Panel A examines whether the association differs between firms in high- and low-litigation industries. The coefficient on CEO-EXEC CONN is statistically insignificant for firms in high-litigation industries (−0.051, t-stat = −0.63), but negative and significant for firms in low-litigation industries (−0.171, t-stat = −3.46). The difference between the two coefficients is statistically significant (p = 0.010). The results using MAJ CEO-EXEC CONN are similar. These findings suggest that the negative association between CEO–executive connections and audit fees is attenuated when litigation exposure is higher, consistent with auditors being more cautious in translating potentially favorable organizational characteristics into lower audit fees in high-litigation settings.

Panel B examines whether the association varies with auditor tenure. The coefficient on CEO-EXEC CONN is negative and significant both when auditor tenure is above the sample median (−0.193, t-stat = −3.68) and when auditor tenure is at or below the sample median (−0.133, t-stat = −2.37). However, the negative association is stronger for longer-tenure auditors, and the difference between the two coefficients is statistically significant at the 10 percent level (p = 0.060). The results using MAJ CEO-EXEC CONN are similar. These findings suggest that CEO–executive connections are more strongly reflected in audit pricing when auditors have greater client-specific knowledge.

Panel C examines whether the association differs between industry specialist and non-specialist auditors. Following prior research on auditor industry specialization, I define IND_EXPERT as an indicator variable equal to 1 if the auditor has the largest audit-fee market share in the client’s two-digit SIC industry-year—based on client firm size measured with total assets—and if its market share exceeds that of the second-largest auditor by at least 10 percentage points, 0 otherwise (Balsam et al., 2003; Mayhew & Wilkins, 2003; Reichelt & Wang, 2010)6. The coefficient on CEO-EXEC CONN is negative and significant for both industry-specialist auditors (−0.327, t-stat = −3.39) and non-specialist auditors (−0.137, t-stat = −2.91). The difference between the two coefficients is statistically significant (p < 0.001), indicating that the negative association is significantly stronger for industry specialists. The results using MAJ CEO-EXEC CONN are similar. These findings suggest that CEO–executive connections are more strongly associated with audit pricing when auditors have greater industry-specific expertise.

Overall, the cross-sectional evidence is consistent with the audit fee association reflecting auditors’ engagement-risk assessments of CEO–executive connections. I note, however, that the tenure and specialization results are also consistent with more experienced and specialized auditors being better able to detect genuine differences in underlying client risk, rather than uniquely reflecting auditors’ use of connections themselves as a risk. As such, while these tests reinforce the plausibility of an auditor risk-assessment interpretation, they do not, on their own, distinguish this interpretation from the alternative that connected top management teams have lower underlying business or financial-reporting risk that all auditors, regardless of experience, would price similarly.

5. Additional Analyses and Robustness Checks

5.1. Types of Connections

The connection measure used in this study encompasses network connections stemming from various activities. However, it is plausible that these diverse connections exert varying effects. The existing literature (e.g., Bruynseels & Cardinaels, 2014; Balsam & Kwack, 2022) acknowledges the non-uniform nature of connections, delineating discernible differences. I follow Bruynseels and Cardinaels (2014) and separate connections (CEO-EXEC CONN or MAJ CEO-EXEC CONN) into those arising from advice networks (CEO-EXEC ADVICE CONN or MAJ CEO-EXEC ADVICE CONN) and friendship networks (CEO-EXEC FRIEND CONN or MAJ CEO-EXEC FRIEND CONN). Advice-based connections are formed through education and employment networks and are more likely to facilitate work-related information sharing and coordination. Friendship-based connections are formed through other activities, such as clubs, charitable organizations, or other social affiliations, and may facilitate discussion of controversial issues. Bruynseels and Cardinaels (2014) find a negative association between CEO–audit committee friendship connections and financial reporting quality, as well as auditor oversight, while no such association is observed for advice connections. This suggests that friendship-based and advice-based connections may have different implications for audit risk and audit pricing.

The results after separating the connection into the two types are reported in

Table 5. Types of Connections.

Dependent Variable	LN(AUDFEE)
CEO-EXEC ADVICE CONN	−0.149 ***
	(−3.39)
CEO-EXEC FRIEND CONN	−0.354
	(−1.15)
Controls	YES
YEAR FE	YES
INDUSTRY FE	YES
Observation	22,220
Adj. R2	0.817

Note: This table shows the OLS regression results on the association between the CEO–executive advice and friendship connections and audit fees. The definitions of the variables are in Appendix A. t-statistics in parentheses are based on clustered standard errors at the firm level. ***, **, and * represent significance levels of 1 percent, 5 percent, and 10 percent, respectively.

. CEO-EXEC ADVICE CONN is negative and significant at the 1 percent level, whereas CEO-EXEC FRIEND CONN is not statistically significant. These results suggest that the negative association between CEO–executive connections and auditors’ pricing is primarily driven by advice-based ties. This is consistent with the view that work-related social ties facilitate information flow, communication, and coordination within the top management team, thereby lowering audit risk or client business risk.

5.2. Endogeneity

Although the main analyses include a broad set of control variables, year fixed effects, and industry fixed effects, the association between CEO–executive connections and audit fees may still be affected by endogeneity concerns. It is possible that executives with connections to the CEO are appointed as top executives of firms that have lower underlying business risk, stronger internal coordination, or higher-quality reporting environments. Also, unobservable firm characteristics may affect both the formation of CEO–executive connections and audit pricing. To address these concerns, I conduct three robustness tests: (1) two-stage instrumental variable estimation, (2) propensity score matched analysis, and (3) propensity score-matched difference-in-differences analysis around CEO turnover.

First, I use a two-stage instrumental variable (IV) estimation. In the first-stage, I estimate the observed level of connections between the CEO and the executives using LN(CEONETWORK) as an instrumental variable, together with the full set of control variables, year fixed effects, and industry fixed effects. LN(CEONETWORK) captures the size of the CEO’s broader social network, excluding ties to individuals within the current firm. As this measure is constructed outside the CEO’s current top management team, it is not mechanically determined by the CEO–executive connections used as the main test variable. I use this measure as an instrument for CEO–executive connections because the CEO’s broader external network captures variation in the CEO’s observable social connectedness that is related to CEO–executive connections but is not based on ties to current executives. Panel A of

Table 6. Robustness Tests.

Panel A. Two-Stage Instrumental Variable Estimation
	1st Stage	2nd Stage
Dependent Variable	CEO-EXEC CONN	LN(AUDFEE)
	Column (1)	Column (2)
LN(CEONETWORK)	−0.007 ***
	(−3.17)
CEO-EXEC CONN		−2.903 **
		(−2.37)
CEO-AC CONN		1.306 **
		(2.21)
Controls	YES	YES
YEAR FE	YES	YES
INDUSTRY FE	YES	YES
Observation	22,220	22,220
Panel B. Propensity Score Matched Sample
Dependent Variable	LN(AUDFEE)
CEO-EXEC CONN	−0.236 ***
	(−4.28)
CEO-AC CONN	−0.020
	(−0.43)
Controls	YES
YEAR FE	YES
INDUSTRY FE	YES
Observation	2908
Panel C. Propensity Score-Matched Difference-in-Differences Analysis around CEO Turnover
Dependent Variable	LN(AUDFEE)
POST	0.274 ***
	(6.87)
POST X TREAT	−0.127 **
	(−2.59)
FIRM FE	YES
Observation	1164

Note: This table shows the results on the association between CEO–executive connections and audit fees using two-stage instrumental variable approach, propensity score matching, and propensity score-matched difference-in-differences analysis around CEO turnover in Panels A, B, and C, respectively. The definitions of the variables are in Appendix A. t-statistics in parentheses are based on clustered standard errors at the firm level. ***, **, and * represent significance levels of 1 percent, 5 percent, and 10 percent, respectively.

reports the results. In the first stage, the coefficient on LN(CEONETWORK) is negative and statistically significant (−0.007, t-stat = −3.17), indicating that the instrument is associated with CEO-EXEC CONN.7 In the second stage, the coefficient on the instrumented value of CEO-EXEC CONN is negative and significant (−2.903, t-stat = −2.37). These results are consistent with the main finding that CEO–executive connections are negatively associated with audit fees.

Second, to make sure that the results documented in this study are not driven by the difference between the firms with a majority of executives with connections to the CEO and the firms with less than a majority of executives with connections to the CEO, I use propensity score-matched (PSM) sample to test the robustness of the results. First, I estimate the conditional odds of having a majority of executives with connections to the CEO using a logistic regression model with MAJ CEO-EXEC CONN as the dependent variable. I include all firm characteristics used in Model (1) as independent variables. Then, I use nearest-neighbor matching within a caliper distance of 0.05 without replacement to match each firm in the treatment group (i.e., MAJ CEO-EXEC CONN = 1) to a firm in the control group (i.e., MAJ CEO-EXEC CONN = 0). This yields 1454 matched pairs or a total of 2908 observations. Although untabulated, post-matching covariate balance tests indicate that 20 out of 22 covariates used in the matching model are statistically indistinguishable between the treatment and control groups. Using the PSM sample of 2908 observations, I re-estimate Model (1) and report the results in Panel B of Table 6. The results are similar to those in Table 3.

Third, I conduct a propensity score-matched difference-in-differences (PSM-DID) analysis around CEO turnover. CEO turnover events are obtained from Gentry et al. (2021), who provide an open-source database of CEO turnover and dismissal events for S&P 1500 firms. CEO turnover provides a setting in which the structure of CEO–executive connections can change following the appointment of a new CEO, while the matched-sample design helps compare firms with pre-turnover CEO–executive connections to observationally similar firms without such connections. I use all CEO turnover events with the required pre- and post-turnover observations because restricting the analysis to plausibly exogenous CEO departures, such as departures due to death or health reasons, leaves too little treatment variation to support reliable estimation.8 I define POST as an indicator variable equal to 1 for the fiscal year immediately after CEO turnover and 0 for the fiscal year immediately before CEO turnover. I exclude the turnover year because CEO–executive connections and audit fees in that year may reflect both the outgoing and incoming CEOs, making the timing of the treatment difficult to interpret. TREAT is an indicator variable equal to 1 if the firm had CEO–executive connections in the pre-turnover year, and 0 otherwise. I estimate the propensity of having pre-turnover CEO–executive connections using a logistic regression model with pre-turnover firm size, measured by LN(ASSET), and profitability, measured by ROA, as independent variables. I then use one-to-one nearest-neighbor matching without replacement to match each treated firm to a control firm based on the estimated propensity score. The DID specification includes firm fixed effects, which control for time-invariant firm characteristics. The coefficient of interest is the interaction term, POST × TREAT, which captures the change in audit fees for treated firms after CEO turnover relative to control firms. As CEO turnover itself may be endogenous to firm performance or governance quality, I do not treat this analysis as a clean identification strategy on its own; rather, it complements the IV and PSM results by exploiting within-firm variation around a salient organizational change. Panel C of Table 6 reports the results using 1164 firm-year observations. The coefficient on POST is positive and significant (0.274, t-stat = 6.87), indicating that audit fees increase after CEO turnover. More importantly, the coefficient on POST × TREAT is negative and significant (−0.127, t-stat = −2.59), suggesting that audit fees increase less for treated firms relative to control firms after CEO turnover. This result presents complementary evidence consistent with the main finding that CEO–executive connections are associated with lower audit fees.

6. Discussion

6.1. Theoretical Implications

The findings contribute to research on organizational social ties by extending its focus beyond the connected individuals themselves. Much of this literature examines outcomes for the connected executives or the firm internally (Granovetter, 1973; Uzzi, 1996; Westphal, 1999). The evidence in this study suggests that social ties within the top management team are also associated with audit pricing, an external market outcome determined by assurance providers. Thus, internal social structures may have implications not only for managerial behavior and internal organizational outcomes, but also for how the organization is assessed by external professionals.

The results also have implications for audit pricing and auditor risk assessment. Auditors evaluate formal indicators such as financial ratios, internal control weaknesses, restatements, and governance structures, but they also consider qualitative information related to management style, tone at the top, and the control environment (PCAOB, 2007; Schmidt, 2014). Recent research on audit data analytics and technology-assisted auditing further suggests that, even as formal risk assessment tools evolve, auditors continue to exercise judgement in evaluating and interpreting risk-relevant information (Commerford et al., 2022; Fedyk et al., 2022). The archival evidence in this study cannot directly observe auditors’ evaluative process or risk perceptions. However, the cross-sectional patterns documented in Section 4.3 are consistent with CEO–executive social ties being reflected in audit pricing in settings where auditor incentives and auditor knowledge make such informal organizational features more or less relevant.

6.2. Limitations

This study has several limitations. First, the archival research design does not directly observe auditors’ risk assessments, or negotiations with clients. Audit fees are observable outcomes, but they do not reveal the specific judgments, information sets, or negotiations underlying audit pricing. Future research could use experiments, surveys, interviews, or audit-workpaper data to examine whether and how auditors consider top management social ties when assessing client risk.

Second, social ties are measured using disclosed information in BoardEx. The measure captures observable connections through employment, education, and other affiliations, but it may miss undisclosed relationships or differences in tie strength. Future research could combine archival data with survey or interview evidence to distinguish weak professional ties from stronger friendship ties (Granovetter, 1973; Bruynseels & Cardinaels, 2014).

Third, although the cross-sectional tests in Section 4.3 are consistent with an auditor risk-assessment interpretation, the evidence does not determine whether connected executive teams are objectively less risky or whether auditors price them as lower risk. Distinguishing between lower underlying audit or client business risk and auditors’ assessment of such risk is an important topic for future research. Studies that combine archival outcomes with direct measures of auditor risk assessments would be especially useful.

Finally, the sample period ends in 2014. This endpoint reflects the availability of the CEO–executive connection data, which require individual-level matching between ExecuComp and BoardEx. As the sample predates critical audit matter (CAM) reporting and more recent developments in audit technology, governance practices, and the audit reporting environment, the findings should be interpreted as evidence from the 2002–2014 institutional setting. However, CAM requirements principally changed how auditors publicly communicate matters involving especially challenging, subjective, or complex judgments. In addition, developments in audit data analytics and technology-assisted auditing may affect how auditors gather, process, and document risk-relevant information. However, these developments do not replace auditors’ underlying responsibility to assess engagement risk, plan the audit, and obtain sufficient appropriate audit evidence. Therefore, although these institutional developments may affect auditors’ documentation, communication, and reporting incentives, I do not believe that the basic association between CEO–executive connections and audit pricing to reverse in the more recent audit environment, although this remains an important question for future research.

7. Conclusions

This study examines whether CEO–executive social ties are associated with audit pricing. Drawing on research on organizational social networks and auditor risk assessment, I argue that CEO–executive connections may have competing implications for audit risk and client business risk. Such ties may facilitate information sharing, coordination, and mutual monitoring within the top management team, thereby reducing audit or client business risk. Alternatively, they may reduce independent challenge or increase collusive potential, thereby increasing risk.

Using S&P 1500 firm-year observations from 2002 to 2014, I find that CEO–executive social ties are negatively associated with audit fees. This evidence is consistent with connected top management teams being associated with lower underlying audit or client business risk, or with auditors pricing such clients as lower risks. Cross-sectional analyses provide additional evidence consistent with this interpretation. I find that litigation risk constrains the audit-fee association, suggesting that litigation risk acts as a boundary condition. In contrast, the association is stronger when auditors have longer tenure and when auditors are industry specialists, suggesting that such connections are more strongly reflected in audit pricing when auditors have greater client-specific or industry-specific knowledge. Overall, the study provides archival evidence that informal social ties within the client’s top management team are associated with audit pricing. The findings contribute to the audit pricing, auditor risk-assessment, and social network literatures. At the same time, because the archival design does not directly observe auditors’ risk perceptions, the results should be interpreted as evidence of an association between CEO–executive connections and audit pricing, rather than as direct evidence of auditors’ judgment processes.