Auditor Appointment Strategy and Intragroup Value Transfers: Evidence From Family Business Groups

Abstract

Family business groups (FBGs) typically control several member firms and can hire a single auditor or multiple auditors to audit their member firms. This article examines what type of auditor appointment strategy constrains intragroup value transfers within FBGs. Analyzing related-party transactions (RPTs) within FBGs in Hong Kong, this study provides evidence that FBGs with multiple auditors undertake more intragroup value transfers than FBGs with a single auditor. However, the adverse effect of multiple-auditor appointments is mitigated by a stronger board and higher financial reporting comparability among member firms. Using an alternative measure of intragroup value transfers, we also find that the market perceives multiple-auditor appointments as impairing audit effectiveness. Overall, our findings offer the new insight that controlling families can exploit the appointment of multiple auditors as a “divide and conquer” strategy which undermines the monitoring role of auditors against intragroup value transfers, but stronger corporate governance of member firms can mitigate the adverse effect.

Keywords

family business groups (FBGs)multiple auditors intragroup value transfers related-party transactions (RPTs)board of directors

Introduction

Family firms play an important role in the world economy. A typical family business group (FBG) consists of several legally independent member firms. Because they operate under the effective control of the same family, intragroup value transfers are pervasive and are often used by controlling families to expropriate minority shareholder wealth, especially in emerging economies (Baek et al., 2006; Bertrand et al., 2002; Friedman et al., 2003; Johnson et al., 2000; Morck et al., 2005).

External auditing can serve as an effective governance mechanism to discipline the expropriation of minority shareholders and mitigate agency conflicts between controlling owners and minority shareholders within FBGs (T. Choi & Wong, 2007; Fan & Wong, 2005; Francis & Wang, 2008). Compared with stand-alone firms, FBGs can enjoy more flexibility and higher bargaining power in selecting external auditors for their member firms, which provides controlling families with a unique opportunity to influence the disciplining role played by auditors. Although some business groups hire a single auditor for all of their member firms, others hire multiple auditors to separately audit different member firms. For example, the Li Ka-Shing conglomerate, the largest FBG in Hong Kong, appointed three Big 4 auditors (Deloitte, PwC, and KPMG) to audit its eight listed member firms as of the end of fiscal year 2010 (see Figure 1). Anecdotal evidence suggests that this appointment of different auditors affects audit efficacy.¹

Figure 1.

The Li Ka-Shing family business group.

This study investigates whether the monitoring role of auditors against intragroup value transfers is affected by FBGs’ auditor appointment strategies (i.e., a single-auditor vs. multiple-auditor appointments). On one hand, a multiple-auditor appointment could weaken the monitoring role of auditors because it can serve as a “divide and conquer” strategy for the controlling families to facilitate intragroup value transfers. Auditors are more reluctant to trust and accept work performed by engagement teams from other audit firms (Joe & Vandervelde, 2007). Therefore, between-firm cooperation and communication among engagement teams from different member firms when multiple auditors are appointed are not as effective as within-firm cooperation and communication among engagement teams from a single auditor (Posner et al., 2010). In the multiple-auditor environment, individual auditors could be denied full access to financial materials of member firms audited by other auditors. This makes it difficult for each auditor to obtain a complete picture of intragroup connectedness and discover intragroup value transfers. Because different auditors develop their own unique audit styles, appointing different auditors may reduce the comparability of financial reporting among member firms (Francis et al., 2014). Lower comparability increases the cost of acquiring information and reduces the overall quantity and quality of information available to market participants (DeFranco et al., 2011). Thus, FBGs that appoint multiple auditors could enjoy greater discretion to undertake self-serving intragroup value transfers.

On the other hand, a multiple-auditor appointment might strengthen the auditor’s ability to monitor against intragroup value transfers. In the multiple-auditor environment, each auditor might have to commit more effort to provide assurance of financial reports. Dual checks by different auditors could discourage controlling families from undertaking intragroup value transfers. Furthermore, it would be more difficult for controlling families to collude to hide any negative effects of intragroup value transfers with multiple auditors than with a single auditor. Auditors also face a lower risk of losing the entire group of clients under the multiple-auditor environment and are less likely to acquiesce to an FBG’s expropriation (J. H. Choi et al., 2010; Nelson et al., 2002).

By analyzing related-party transactions (RPTs) of member firms within FBGs in Hong Kong from 2003 through 2017, this study provides evidence regarding the effect of auditor appointment strategies on external auditors’ monitoring of intragroup value transfers. Hong Kong provides an ideal research setting to investigate our research question for several reasons. First, FBGs are prevalent in Hong Kong, providing a sufficient sample for our empirical analyses. Second, controlling families in Hong Kong are actively involved in running their group firms and rarely delegate strategic and resource allocation decisions (such as auditor choice) to professional managers (Carney, 2008). Group-affiliated firms in Hong Kong are thus vulnerable to intragroup value transfers by controlling families. Third, prior literature documents that practice offices are an appropriate unit of analysis in audit research (Francis et al., 1999). The audit market in Hong Kong is basically one large city-level market because audits are performed out of one central office rather than multiple offices throughout the territory (DeFond et al., 2000).

RPTs serve as a direct channel for controlling shareholders to transfer value in or out of their member firms for private benefits (Chang & Hong, 2000; Cheung et al., 2006; Jian & Wong, 2010).² We find that the incidence of RPTs is significantly higher within FBGs that hire multiple auditors than within FBGs that hire a single auditor. Holding all else equal, the frequency of RPTs in FBGs with multiple auditors is 24.61% higher than that in FBGs with a single auditor. This adverse effect of multiple-auditor appointments is not fully driven by the involvement of a non-Big 4 auditor among the multiple auditors.

Our additional analyses show that the adverse effect of multiple-auditor appointments on intragroup value transfers is significantly mitigated when member firms have more independent boards but is aggravated when a larger number of member firms have a controlling family member serving as board chairperson. The adverse effect of multiple-auditor appointments is also attenuated by higher accounting comparability among member firms.

Finally, we conduct an analysis that is complementary to the RPT tests. For an alternative measure of intragroup value transfers, we rely on G. S. Bae et al. (2008) and examine market reactions of member firms to earnings news released by their group peers.³ Our results show that the positive market reactions of member firms to earnings announcements made by group peers are more pronounced when FBGs appoint multiple auditors rather than a single auditor. These results suggest that the market perceives multiple-auditor appointments as impairing audit effectiveness in disciplining intragroup value transfers within FBGs.

This study contributes to the literature as follows. First, it identifies a unique research setting within FBGs and contributes to the auditing literature beyond studies on the choice of a Big N versus a non-Big N auditor in mitigating agency conflicts between controlling owners and minority shareholders (Fan & Wong, 2005; Fang et al., 2017; Haw et al., 2011; Kim & Yi, 2006). Whereas prior studies typically find that hiring Big N auditors improves disclosure and earnings quality, this study finds that the appointment of different auditors within an FBG exacerbates agency problems related to the controlling family ownership.

Second, our study contributes to the corporate governance literature (Byrd & Hickman, 1992; Cheung et al., 2014; Gillette et al., 2003) by showing that stronger internal corporate governance of member firms can mitigate the adverse impact of multiple-auditor appointments within FBGs. This evidence sheds new light on an effective monitoring mechanism within FBGs, and offers valuable implications to regulators and policy makers, especially in East Asian economies where FBGs are pervasive and typical Type II agency problems are severe.⁴

The rest of this article proceeds as follows. The section “Literature Review and Hypothesis Development” develops our research hypothesis. The section “Research Design” presents the research design. The section “Sample Selection” describes our sample selection. The section “Empirical Results” reports empirical results, and the section “Conclusion” concludes.

Literature Review and Hypothesis Development

Intragroup Value Transfers Among Group Member Firms

An FBG typically comprised a set of member firms controlled by the common family. One justification for the prevalence of FBGs is that member firm resources could be transferred to bail out financially distressed group peers (G. S. Bae et al., 2008; Friedman et al., 2003; Riyanto & Toolsema, 2008) or to ease financial constraints of group peers with profitable investment opportunities (Chang & Hong, 2000; Shin & Park, 1999).

However, intragroup value transfers can also be exploited by controlling families to expropriate wealth from minority shareholders of member firms (Johnson et al., 2000; Morck & Yeung, 2003; Morck et al., 2005). Typically, controlling families tend to transfer resources from their lower- to higher-cash-flow-rights member firms (K. Bae et al., 2002; Baek et al., 2006; Bertrand et al., 2002). The interests of member firms’ minority shareholders will be impaired if the member firm resources are misappropriated.

Auditor Appointment Strategy and Intragroup Value Transfers

External auditors perform an important governance role in deterring the controlling shareholders’ misappropriation of firm resources (Fan & Wong, 2005; Fang et al., 2017; Haw et al., 2011; Kim & Yi, 2006). Unlike stand-alone firms, FBGs could exploit the single-auditor versus multiple-auditor appointment strategy to affect the effectiveness of audit work. Existing literature suggests two conflicting predictions regarding the potential impact of the auditor appointment strategy on auditor monitoring against intragroup value transfers within FBGs. We discuss these next.

Divide and conquer strategy

Controlling families could exploit the multiple-auditor appointment as a “divide and conquer” strategy to undermine external auditors’ monitoring of intragroup value transfers. By appointing multiple auditors, FBGs assign different auditors to perform audit tasks for different member firms, preventing each individual auditor from having full access to the financial materials of member firms audited by other auditors. On one hand, this would make it more difficult for each auditor to evaluate the implications of intragroup connectedness for audit risk, and thus to effectively detect controlling families’ opportunistic use of intragroup value transfers to manipulate earnings. On the other hand, this divide and conquer strategy could distort the communication and cooperation between engagement teams from different audit firms. Because auditors are more reluctant to trust and accept work performed by engagement teams from other audit firms than from the same audit firm (Joe & Vandervelde, 2007; King, 2002; Tajfel & Turner, 1986), the communication and cooperation between engagement teams from different auditors are rarely as effective as between teams from the same auditor. The poorer communication and coordination between different auditors could weaken each auditor’s ability to monitor against intragroup value transfers within FBGs.

Different auditors have their own unique style, which lowers the comparability of financial statements among firms audited by different auditors (Francis et al., 2014). Financial reporting with lower comparability increases the cost of acquiring information and reduces the overall quantity and quality of information available to outsiders (DeFranco et al., 2011). Therefore, reduced accounting comparability arising from multiple-auditor appointments could impede auditors and investors from seeing through the negative consequences of intragroup value transfers on financial reports and from taking timely actions against intragroup value transfers.

Finally, member firms within FBGs may exhibit substantial similarity in corporate behaviors. Therefore, knowledge about group-wide business conditions and internal control effectiveness obtained through auditing member firms can be used to improve the efficiency and effectiveness of group peers’ audits. However, multiple-auditor appointments sever knowledge spillovers impairing audit effectiveness in constraining intragroup value transfers.

Cross-monitoring view

In a typical business group, the pervasive operational and financial connectedness between member firms might increase audit risk faced by auditors. Contrary to the divide and conquer view, a multiple-auditor appointment might strengthen cross-monitoring of value transfer transactions between member firms if each auditor exercises a heightened degree of professional skepticism and becomes more diligent in auditing intragroup connectedness. Specifically, each audit firm might assign more experienced engagement team personnel, apply greater engagement effort, perform auditing procedures more prudently, and lower the threshold for material errors to reduce increased audit risk (Bedard & Johnstone, 2004; Bell et al., 2001, 2008). This extensive audit efforts could deter member firms from managing earnings via intragroup transactions and increase the possibility of detecting any earnings management to conceal intragroup value transfers.

Controlling families are likely to have more power under the single-auditor appointment than the multiple-auditor appointment because in a single-auditor appointment, the auditor faces the risk of losing all member firm clients in the case of disagreements with the controlling family. In contrast, the lower economic dependence of each individual auditor on FBGs under the multiple-auditor appointment promotes auditor independence and renders each auditor less likely to acquiesce to the pressures imposed by the controlling family (J. H. Choi et al., 2010; Nelson et al., 2002). It is thus more difficult for the controlling family to collude with multiple auditors than with a single auditor to manipulate earnings via intragroup value transfers or to conceal intragroup resources diversion via earnings management.

RPTs to Capture Intragroup Value Transfers

RPTs are a direct way for the controlling family to transfer resources in or out of their group member firms. The controlling family can undertake RPTs to divert firm resources and expropriate wealth at the expense of minority shareholders (Djankov et al., 2008; Johnson et al., 2000). For example, the controlling families of Korean chaebols conduct intragroup deals (particularly, acquisitions and equity-linked private securities offerings) to increase their wealth at the expense of minority shareholders of member firms (K. Bae et al., 2002; Baek et al., 2006).

Auditors should pay attention to RPTs because RPTs are associated with earnings manipulation and the expropriation of minority shareholders (Fisman & Wang, 2010; Kohlbeck & Mayhew, 2017; Lo et al., 2010). Auditors tend to exert additional audit effort to investigate opportunistic RPTs (Habib et al., 2015; Kohlbeck & Mayhew, 2017), and then are more inclined to issue modified audit opinions to communicate the risks associated with RPTs to capital market participants (Fang et al., 2018). Therefore, auditors, especially reputable ones, can play an important governance role in monitoring RPTs (Bennouri et al., 2015; Cheung et al., 2006).

If the auditor appointment strategy affects the auditor’s monitoring against intragroup value transfers within FBGs, we predict that the incidence of RPTs will be significantly different between FBGs that hire a single auditor and FBGs that hire multiple auditors. Given that it is ex ante unknown whether the cross-monitoring view or the divide and conquer motive associated with a single- versus multiple-auditor appointment prevails, we state the following null hypothesis:

Hypothesis (H1): There is no difference in the incidence of RPTs between FBGs that hire a single auditor and FBGs that hire multiple auditors for their member firms.

Research Design

We specify the following regression model (Equation 1) to examine the impact of the auditor appointment strategy on the frequency with which FBGs undertake intragroup value transfers via RPTs:

\begin{matrix} LOGRPTFRE Q_{it} = α_{0} + β_{1} MUL T_{it} / - 1 HH I_{it} + β_{2} GWEDG E_{it} + \\ β_{3} LnNFIRM S_{it} + β_{4} LnNSECTOR S_{it} + β_{5} LnBLOCK S_{it} + \\ β_{6} LnGBODSIZ E_{it} + β_{7} GINDDI R_{it} + β_{8} GFMLYCHAI R_{it} + \\ β_{9} GSIZ E_{it - 1} + β_{10} GLE V_{it - 1} + β_{11} GM B_{it - 1} + β_{12} GRO A_{it - 1} + \\ β_{13} LnGAUDFE E_{it - 1} + β_{14} GACCTCOM P_{it - 1} + Year dummies + ε, \end{matrix}

(1)

where the dependent variable, LOGRPTFREQ_it, is the natural log of one plus the number of RPTs undertaken by the FBG in year t. Our variable of interest, MULT/−1HHI, represents two alternative measures of FBGs’ auditor appointment strategies. MULT is a dummy variable equal to 1 if an FBG hires multiple auditors in fiscal year t, and 0 otherwise. −1HHI is the Herfindahl-Hirschman concentration index per group, multiplied by −1. We measure the Herfindahl-Hirschman index based on the total number of client firms audited by each auditor in a group-year (Boone et al., 2011; Francis et al., 2013).⁵ The coefficients on MULT and –1HHI will be positive (negative) if the divide and conquer motive (cross-monitoring view) of multiple-auditor appointments dominates.

We include control variables from prior studies (Bennouri et al., 2015; Cheung et al., 2006) in Equation 1. We include GWEDGE to capture the effect of the difference between the control rights and cash flow rights of the controlling owners. This is positively related to expropriation of minority shareholders (Claessens et al., 2000, 2002). Following prior literature (Fan & Wong, 2002; Haw et al., 2004), we define the divergence of control rights from cash flow rights as 1 minus cash flow rights divided by control rights, and measure GWEDGE as the median divergence of member firms within the FBG.⁶

A high quality board of directors can better protect the interests of minority shareholders and constrain intragroup value transfers (Cheung et al., 2014). We thus include board size (LnGBODSIZE), board independence (GINDDIR), and board leadership (GFMLYCHAIR) to proxy for board quality. LnGBODSIZE is the natural log of the median number of directors in member firms, GINDDIR is the median percentage of independent directors in the FBG’s member firms, and GFMLYCHAIR is the fraction of member firms within the FBG that have a controlling family member serving as board chairperson.

The more member firms that an FBG has, the more likely it is to undertake RPTs. We thus include LnNFIRMS, defined as the log of the number of member firms within an FBG. The variable LnBLOCKS is the natural log of one plus the number of member firms with at least one blockholder holding no less than 5% cash flow rights within the group. We include LnNSECTORS, measured as the log of the number of industries in which the FBG’s member firms operate. We further control for the group size (GSIZE), the group-level financial leverage (GLEV), market-to-book ratio (GMB), return on assets (GROA), audit fees paid to external auditors (LnGAUDFEE), and accounting comparability (GACCTCOMP) in the previous fiscal year. We define GACCTCOMP as the median value of the firm-year measure of accounting comparability (ACCTCOMP) of all member firms in the year.

Following DeFranco et al. (2011), we measure financial statement comparability by estimating the following model for each firm-year using data from 16 quarters prior to fiscal year end:⁷

Earning s_{it} = α_{i} + β_{i} Retur n_{it} + ε_{it} .

(2)

We produce ACCTCOMP by aggregating accounting comparability of all pairs between member firm i and its group peers in the year.⁸ Detailed definitions of variables used in Equation 1 are presented in the appendix. Finally, we include year dummies to control for time effects.⁹

Sample Selection

Our sample consists of FBGs and their listed member firms in Hong Kong from 2003 through 2017. Specifically, we identify FBGs and their listed member firms by manually collecting ownership data from annual reports of all firms listed on the Hong Kong Stock Exchange (Main Board), and then grouping listed firms that share the same ultimate controlling owner into the same FBG.¹⁰ We require that each FBG has no less than two listed member firms, and member firms did not enter their business groups through mergers and acquisitions or spin-offs.¹¹ Initially, we obtain 171 member firms affiliated with 59 FBGs.

We manually collect data on RPTs from the HKExnews website, an information disclosure archive for listed firms launched by the Hong Kong Stock Exchange.^12,13 We follow Cheung et al. (2006) to select opportunistic RPTs that occurred between member firms and their controlling families or controlling families’ related parties. Those RPTs include acquisitions (sales) of assets by member firms from (to) related parties, sales of equity stake in member firms to related parties, trading relationships between member firms and related parties, and direct cash payments to related parties. Data on auditor identity, ownership structure, and the board of directors are hand-collected from the sample member firm’s annual reports. We extract financial data from the DataStream database. We obtain 726 group-year observations involving 705 RPTs undertaken by 167 member firms within 57 FBGs.

Panel A in Table 1 reports the distribution of RPTs across FBGs with different auditor appointment strategies. Among the 705 RPTs, 357 (24) come from FBGs that hire a single Big 4 (a non-Big 4) auditor; 165 come from FBGs that hire multiple Big 4 auditors; and 159 come from FBGs that hire both Big 4 and non-Big 4 auditors. Panel B reports the sample distribution across auditors. Deloitte is engaged in the largest fraction of group-year observations, followed by PwC, KPMG, and EY. Panel C presents the distribution of sample member firms across industries based on the two-digit-Standard-Industrial-Classification (SIC) codes. Our sample member firms mainly come from service, construction, real estate, manufacturing, and finance industries.¹⁴

Table 1.

Sample Distribution Across FBGs With Different Type of Auditor Appointment.

Panel A: Sample Distribution Across Group Auditor Appointment Types.
Group auditor appointment	Group-years	Groups	RPTs
One Big 4	493	46	357
One non-Big 4	36	4	24
Multiple Big 4	127	17	165
Big 4 and non-Big 4	70	9	159
Total	726	76	705

Panel B: Sample Distribution Across Auditors.
Auditor	Group-years	Member firms	Family business groups
Deloitte	358	74	35
PwC	225	51	24
EY	129	33	15
KPMG	169	26	13
Non-Big 4	112	29	20
Total	993	213	107

Panel C: Sample Distribution Across Industries at Member Firm Level.
Industry	Two-digit-SIC-Code	No. of firms
Mining	12	1
Construction	15, 16, 17	28
Manufacturing	20, 23, 26, 27, 28, 30, 32, 35, 36	24
Transportation	41, 42, 44, 45, 47	7
Communications	48	5
Electric and gas	49	2
Wholesale trade	50, 51	6
Retail trade	53, 56, 58, 59	11
Finance	60, 61, 62, 67	20
Real estate	65	26
Service	70, 73, 78, 79, 80, 82, 87	37
Total		167

Note. This table reports the distribution of sample across FBGs with different auditor appointment strategies. The sample comprised 726 group-year observations involving 705 RPT announcements made by 57 FBGs from 2003 through 2017. FBG = family business group; RPT = related-party transaction.

Empirical Results

Descriptive Statistics

Panel A in Table 2 provides descriptive statistics for the variables used in regression analyses. Overall, 27.1% of our group-year observations come from FBGs that hire multiple auditors (MULT). On average, our sample FBGs undertake 0.971 RPT per year (RPTFREQ). The average number of member firms within an FBG (NFIRMS) is 2.9, of which 1.45 member firms have a large shareholder other than their controlling families (NBLOCKS). The median number of directors on member firms’ boards (GBODSIZE) is 9.17, of which 38.9% are independent nonexecutive directors (GINDDIR). On average, 56.5% of member firms within an FBG have a controlling family member serving as board chairperson.

Table 2.

Summary Statistics and Correlations.

Panel A: Descriptive Statistics.
Variable	N	M	SD	Q1	Median	Q3
RPTFREQ	726	0.971	1.375	0	0	1
LOGRPTFREQ	726	0.498	0.568	0	0	0.693
MULT	726	0.271	0.445	0	0	1
–1HHI	726	–0.868	0.221	–1	–1	–0.625
GWEDGE	726	0.246	0.2	0.125	0.215	0.358
LnNFIRMS	726	0.986	0.37	0.693	0.693	1.099
LnNSECTORS	726	0.552	0.481	0	0.693	0.693
LnNBLOCKS	726	0.758	0.53	0	0.693	1.099
LnGBODSIZE	726	2.179	0.248	2.072	2.197	2.303
GINDDIR	726	0.389	0.091	0.333	0.375	0.429
GFMLYCHAIR	726	0.565	0.371	0.333	0.5	1
GSIZE	726	14.641	1.609	13.546	14.494	15.722
GLEV	726	0.101	0.087	0.031	0.084	0.151
GMB	726	0.937	0.828	0.419	0.7	1.198
GROA	726	0.096	0.192	0.019	0.053	0.099
LnGAUDFEE	726	7.988	0.784	7.495	8.059	8.494
GACCTCOMP	726	0.242	0.362	0.068	0.149	0.263

Panel B: Univariate Tests.
Variable	N	MULT = 0		N	MULT = 1		Difference
		M	SD		M	SD	MULT 1 – 0
RPTFREQ	529	0.720	1.093	197	1.645	1.774	0.925***
LOGRPTFREQ	529	0.395	0.509	197	0.775	0.623	0.380***
GWEDGE	529	0.201	0.165	197	0.368	0.231	0.167***
LnNFIRMS	529	0.874	0.272	197	1.289	0.425	0.415***
LnNSECTORS	529	0.416	0.447	197	0.918	0.366	0.502***
LnNBLOCKS	529	0.691	0.502	197	0.935	0.563	0.244***
LnGBODSIZE	529	2.143	0.227	197	2.277	0.274	0.134***
GINDDIR	529	0.396	0.090	197	0.369	0.088	−0.027***
GFMLYCHAIR	529	0.608	0.368	197	0.451	0.355	–0.157***
GSIZE	529	14.429	1.524	197	15.210	1.695	0.781***
GLEV	529	0.102	0.091	197	0.099	0.074	–0.003
GMB	529	0.930	0.868	197	0.957	0.711	0.027
GROA	529	0.112	0.22	197	0.052	0.064	–0.060***
LnGAUDFEE	529	7.992	0.771	197	7.976	0.82	–0.016
GACCTCOMP	529	0.264	0.411	197	0.184	0.158	–0.080***

Panel C: Pearson Correlations.
		A	B	C	D	E	F	G	H	I	J	K	L	M	N	O
LOGRPTFREQ	A	1
MULT	B	.298***	1
–1HHI	C	.310***	.978***	1
GWEDGE	D	.274***	.372***	.380***	1
LnNFIRMS	E	.483***	.499***	.519***	.560***	1
LnNSECTORS	F	.324***	.464***	.488***	.310***	.659***	1
LnNBLOCKS	G	.300***	.205***	.195***	.251***	.501***	.359***	1
LnGBODSIZE	H	.043	.240***	.277***	.062*	.151***	.080**	.076**	1
GINDDIR	I	–.080**	–.133***	–.148***	–.001	–.130***	–.076**	–.081**	–.661***	1
GFMLYCHAIR	J	–.058	–.188***	–.191***	–.011	–.087**	–.244***	–.206***	–.249***	.108***	1
GSIZE	K	.062*	.216***	.236***	.004	.210***	.111***	.205***	.482***	–.119***	–.195***	1
GLEV	L	.150***	–.012	.002	–.051	.083**	.020	.074**	.107***	–.046	–.106***	.317***	1
GMB	M	.006	.014	.033	–.214***	–.147***	–.050	.008	.043	.000	–.110***	.258***	.032	1
GROA	N	–.064*	–.138***	–.135***	–.129***	–.155***	–.144***	–.039	–.054	–.003	–.012	.110***	.054	.280***	1
LnGAUDFEE	O	.074**	–.009	.002	.028	.141***	.026**	.287***	.330***	–.039	–.233***	.621***	.308***	.079**	.010	1
GACCTCOMP	P	.135***	–.098***	–.105***	.074**	.058	.018	.175***	–.236***	.007	–.122***	–.337***	.023	–.090***	.009	–.232***

Note. This table reports the descriptive statistics, univariate tests, and Pearson correlations. Panel A presents descriptive statistics, Panel B reports univariate comparison, and Panel C reports Pearson correlations. The sample comprised 726 group-year observations involving 705 RPT announcements made by 57 FBGs from 2003 through 2017. All variables are defined in the appendix. FBG = family business group; RPT = related-party transaction.

Denote significance at the 10% level.

Denote significance at the 5% level.

***

Denote significance at the 1% level.

Panel B in Table 2 presents univariate comparisons between FBGs with multiple-auditor appointments and FBGs with a single-auditor appointment. On average, FBGs with multiple-auditor appointments undertake more RPTs (1.645 vs. 0.720) than their counterparts. FBGs with multiple-auditor appointments have significantly higher values of GWEDGE, LnNFIRMS, LnNSECTORS, LnNBLOCKS, LnGBODSIZE, GINDDIR, and GSIZE, but significantly lower values of GFMLYCHAIR, GROA, and GACCTCOMP.

Panel C reports the Pearson correlations among the main variables. The frequency of RPTs (LOGRPTFREQ) is positively and significantly correlated with both MULT and –1HHI, providing preliminary evidence that FBGs with multiple auditors tend to undertake RPTs more frequently than FBGs with a single auditor. In addition, the frequency of RPTs is also positively correlated with the divergence of control rights from cash flow rights (GWEDGE), the number of member firms within an FBG (LnNFIRMS), the number of industries involving an FBG (LnNSECTORS), the number of member firms with blockholders other than the controlling families (LnNBLOCKS), group size (GSIZE), financial leverage ratio (GLEV), audit fees (LnGAUDFEE), and group accounting comparability (GACCTCOMP). However, it is negatively correlated with the percentage of independent directors (GINDDIR) and member firms’ financial performance (GROA). It is interesting to note that group accounting comparability (GACCTCOMP) is negatively correlated with both MULT and –1HHI. This is consistent with the Francis et al.’s (2014) finding that earnings are less comparable among firms audited by different auditors.

Baseline Regression Results

Table 3 reports the regression results examining the impact of multiple-auditor appointments on the frequency of RPTs within FBGs. The results using MULT and –1HHI are presented in Columns 1 and 3, respectively. Columns 2 and 4 test for multicollinearity and show that none of the variance inflation factors (VIFs) are higher than 3.15, suggesting little multicollinearity concern. The regression results show that the coefficients on both MULT and –1HHI are positive and statistically significant, indicating that FBGs with multiple-auditor appointments tend to undertake more RPTs. This impact of multiple-auditor appointments is economically significant as well. Based on the results using MULT in column 1, FBGs that hire multiple auditors conduct 24.61%, exp(0.22) – 1, more RPTs, all else equal.

Table 3.

The Effect of Auditor Appointment Strategy on the Frequency of Related-Party Transactions (RPTs).

Dependent variable = LOGRPTFREQ	(1)	(2)	(3)	(4)
		VIFs		VIFs
MULT	0.220***	1.72
	(3.27)
–1HHI			0.490***	1.85
			(3.30)
GWEDGE	–0.070	1.68	–0.079	1.68
	(−0.47)		(−0.53)
LnNFIRMS	0.586***	3.11	0.574***	3.14
	(3.77)		(3.94)
LnNSECTORS	–0.013	2.04	–0.023	2.07
	(−0.15)		(−0.27)
LnNBLOCKS	0.077	1.64	0.083	1.64
	(1.19)		(1.28)
LnGBODSIZE	–0.150	3.02	–0.180	3.06
	(–0.94)		(−1.12)
GINDDIR	0.256	2.17	0.287	2.17
	(0.77)		(0.88)
GFMLYCHAIR	0.073	1.33	0.072	1.32
	(0.96)		(0.97)
GSIZE	–0.050**	2.68	–0.051**	2.68
	(−2.14)		(−2.17)
GLEV	0.897***	1.25	0.891***	1.25
	(3.37)		(3.37)
GMB	0.036	1.40	0.032	1.41
	(1.15)		(1.04)
GROA	–0.094	1.19	–0.093	1.19
	(−1.06)		(−1.05)
LnGAUDFEE	0.106**	2.39	0.109**	2.39
	(2.34)		(2.47)
GACCTCOMP	0.146**	1.42	0.147**	1.42
	(2.48)		(2.50)
Constant	–0.499*		0.039
	(–1.72)		(0.11)
Year dummies	Yes		Yes
Observations	726		726
Adjusted R²	.316		.319

Note. This table tests whether the incidence of RPTs is lower or higher within FBGs that hire multiple auditors than within FBGs that hire a single auditor. The dependent variable, LOGRPTFREQ, is defined as the natural log of one plus the number of related-party transactions undertaken by member firms within an FBG in fiscal year t. All variables are defined in the appendix. The t statistics given in parenthesis below each estimate are based on robust standard errors clustering at a group level. VIF = variance inflation factor; RPT = related-party transaction; FBG = family business group.

Denote significance at the 10% level.

Denote significance at the 5% level.

***

Denote significance at the 1% level.

The results regarding some control variables are also noteworthy. The frequency of RPTs is positively associated with the number of member firms (LnNFIRMS), group leverage (GLEV), and audit fees (LnGAUDFEE), consistent with findings in prior research that FBGs tend to bail out financially constrained member firms via RPTs (G. S. Bae et al., 2008; Friedman et al., 2003; Gopalan et al., 2014) and that additional audit effort is demanded to reduce higher audit risk associated with RPTs (Habib et al., 2015; Kohlbeck & Mayhew, 2017).

Endogeneity Issues

Both the group auditor appointment and intragroup value transfers are likely to be endogenous because controlling families with a higher propensity to undertake intragroup value transfers could hire multiple auditors to weaken the monitoring from external auditors. To address this potential endogeneity concern, we employ an instrumental variables approach and estimate two-stage least squares (2SLS) regressions. We use the dispersion in foreign sales of member firms within an FBG (DISFSALES) as an instrument for the FBGs’ auditor appointment. Prior studies find that a firm’s foreign sales play a role in determining its auditor choice (Guedhami et al., 2014; Kim et al., 2019). Foreign customers demand differential auditor monitoring over suppliers than domestic customers. However, no theory posits that foreign sales have a direct impact on a firm’s RPTs.

We define the dispersion in foreign sales (DISFSALES) within an FBG as the standard deviation of a member firm’s ratio of foreign sales to total sales within an FBG in year t− 1. In the first stage, we regress the auditor appointment variable on DISFSALES along with other control variables. In the second stage, we regress the frequency of RPTs (LOGRPTFREQ) on the predicted value of the multiple-auditor appointment variable (MULT_HAT or –1HHI_HAT) from the first-stage regression, along with all control variables.

We report the results regarding MULT in Columns 1 and 2, and the results regarding –1HHI in Columns 3 and 4 of Table 4. The results from the first-stage regression are reported in Columns 1 and 3. They show that the coefficient on DISFSALES is positive and statistically significant as expected, consistent with the conjecture that FBGs with a higher dispersion in foreign sales of their member firms are more likely to hire multiple auditors. The explanatory power (R²) of the first-stage regression is relatively high (0.340 and 0.445 in Columns 1 and 3, respectively), suggesting a good fit. The statistics reported in the bottom rows show that the partial F statistic from the first-stage regression for the instrument passes the weak identification tests, and the Kleibergen-Paap LM Statistic also easily passes the associated under-identification tests, suggesting that the chosen instrument is valid (Wooldridge, 2016).

Table 4.

The Effect of Auditor Appointment Strategy on the Frequency of RPTs: 2SLS Analysis.

Dependent variable =	MULT	LOGRPTFREQ	–1HHI	LOGRPTFREQ
	(1)	(2)	(3)	(4)
	First stage	Second stage	First stage	Second stage
DISFSALES	0.202***		0.113***
	(2.58)		(3.04)
MULT_HAT		2.445**
		(2.39)
–1HHI_HAT				4.375***
				(2.70)
Controls	Yes	Yes	Yes	Yes
Year dummies	Yes	Yes	Yes	Yes
Observations	726	726	726	726
Adjusted R²	.340		.445
First-stage partial F statistic	6.65**		9.26***
First-stage partial R²	.0084		.0114
Kleibergen-Paap LM Statistic	7.01***		9.64***

Note. This table reports results from 2SLS regressions. We estimate OLS regression using MULT (–1HHI) as the dependent variable in the first stage, and then include the fitted value of MULT (–1HHI) in place of MULT (–1HHI) in the second-stage regression. The instrumental variable, DISFSALES, is defined as the standard deviation of the member firm’s ratio of foreign sales to total sales within an FBG in year t – 1. The dependent variable in the second-stage OLS regression is LOGRPTFREQ, defined as the natural log of one plus the number of related-party transactions undertaken by member firms within an FBG in fiscal year t. The other variables are defined in the appendix. Each regression includes a separate intercept and the same set of control variables as in Table 3. The t statistics given in parenthesis below each estimate are based on robust standard errors. RPT = related-party transaction; SLS = stage least square; OLS = ordinary least squares; FBG = family business group.

Denote significance at the 5% level.

***

Denote significance at the 1% level.

The results from the second-stage regressions in Columns 2 and 4 indicate that the coefficients on MULT_HAT and –1HHI_HAT remain positive and statistically significant. These findings suggest that our main results in Table 3 are not driven by endogeneity.

The Type of Multiple-Auditor Appointment

Prior literature documents that Big 4 auditors perform a better governance role than non-Big 4 auditors (e.g., Becker et al., 1998; T. Choi & Wong, 2007). Big 4 auditors are also able to more effectively monitor RPTs (Bennouri et al., 2015; Cheung et al., 2006). Therefore, it may be possible that the impact of multiple-auditor appointments on RPTs is driven by the presence of non-Big 4 auditors.

To test this possibility, we separate the variable MULT into two variables: MULTBIG4 indicating multiple-auditor appointments of only Big 4 auditors, and BIG4&non-BIG4 indicating multiple-auditor appointments of both Big 4 and non-Big 4 auditors. We re-estimate Column 1 of Table 3 with these two indicators in place of MULT. The results reported in Column 1 of Table 5 show that the coefficients on both MULTBIG4 and BIG4&non-BIG4 are positive and statistically significant. The difference in the two coefficients is statistically insignificant, as shown by the F statistic in the bottom row. These results do not suggest a significant difference in the adverse effect between the two types of multiple-auditor appointments.

Table 5.

The Effect of Auditor Appointment Strategy on the Frequency of RPTs: Different Types of the Multiple-Auditor Appointment.

Variable	(1)	(2)
MULTBIG4	0.156*
	(1.85)
BIG4&non-BIG4	0.315**
	(2.65)
MULT_PURE		0.367***
		(2.90)
MULT_MIX		0.136*
		(1.86)
Controls	Yes	Yes
Year dummies	Yes	Yes
Observations	726	726
Adjusted R²	.319	.320
Difference in coefficients on BIG4&non-BIG4 and MULTBIG4	0.159
F statistic	1.08
Difference in coefficients on MULT_PURE and MULT_MIX		0.231
F statistic		1.58

Note. This table reports the results of the regression decomposing the variable MULT into different types. The dependent variable, LOGRPTFREQ, is defined as the natural log of one plus the number of related-party transactions undertaken by member firms within an FBG in fiscal year t. The variable, MULTBIG4, is defined as a dummy variable that equals 1 if an FBG hires multiple Big 4 auditors, and 0 otherwise. The variable, BIG4&non-BIG4, is defined as a dummy variable that equals 1 if an FBG hires both Big 4 and non-Big 4 auditors, and 0 otherwise. The variable, MULT_PURE, is defined as a dummy variable that equals 1 if each auditor audits only one member firm within an FBG that hires multiple auditors, and 0 otherwise. The variable, MULT_MIX, is defined as a dummy variable that equals 1 if two or more member firms are audited by the same auditor within an FBG that hires multiple auditors, and 0 otherwise. The other variables are defined in the appendix. Each regression includes a separate intercept and the same set of control variables as in Table 3. The t statistics given in parenthesis below each estimate are based on robust standard errors clustering at a group level. FBG = family business group.

Denote significance at the 10% level.

Denote significance at the 5% level.

***

Denote significance at the 1% level.

Next, we further refine our tests on MULT. Even in the multiple-auditor appointment environment, some firm pairs can have the same auditor. As shown in Figure 1, Li Ka-Shing FBG is characterized as having a multiple-auditor appointment. However, four of eight group member firms are audited by the same auditor (PwC). Such a group could be different from a group of four member firms, each hiring a different Big 4. We thus decompose the variable MULT into two variables: MULT_PURE indicating a multiple-auditor appointment where no member firms are audited by the same auditor, and MULT_MIX indicating a multiple-auditor appointment where two or more member firms are audited by the same auditor.

We re-estimate Column 1 of Table 3 with these two indicators in place of MULT. The results reported in Column 2 of Table 5 show that the coefficients on both MULT_PURE and MULT_MIX are significantly positive. The difference in the two coefficients is not statistically significant, as shown by the F statistic in the bottom row. These results indicate that the observed impact on RPTs is not significantly different between the two types of multiple-auditor appointments.¹⁵

The Role of a Board of Directors

A board of directors is the highest internal governance mechanism responsible for curbing agency conflicts between a firm’s insiders and outsiders (Fama & Jensen, 1983). In this section, we examine whether the impact of the auditor appointment strategy on RPTs varies with board structure. We employ three previously defined board variables (i.e., LnGBODSIZE, GINDDIR, and GFMLYCHAIR) to capture board structure. Larger boards are preferable for firms with greater agency problems because they are able to commit more effort to overseeing insiders (Harris & Raviv, 2008; Raheja, 2005). Independent directors strengthen the viability of a board as an effective internal control mechanism (Dahya et al., 2008; Fama & Jensen, 1983; Gillette et al., 2003). Among board members, the board chairperson plays a more influential role than other directors in corporate governance activities, so controlling families securing the position of board chairperson can weaken the board’s ability to discipline controlling family’s self-serving behaviors (Cheung et al., 2014).

We include the interactions between each of the three board variables and each of MULT and –1HHI in Equation 1 and report the results in Columns 1 and 2 of Panel A in Table 6. The coefficients on the interactions MULT×GINDDIR and –1HHI×GINDDIR are significantly negative whereas the coefficients on the interactions MULT×GFMLYCHAIR and –1HHI×GFMLYCHAIR are significantly positive. These results indicate that more independent boards can mitigate, but boards chaired by the controlling family member aggravate, the impact of multiple-auditor appointments on intragroup value transfers. However, the interactions with board size are not statistically significant.

Table 6.

The Effect of Auditor Appointment Strategy on the Frequency of RPTs: Board Characteristics and Financial Statement Comparability.

Panel A: Board Characteristics.
Variable	(1)	(2)
MULT	0.373
	(0.77)
MULT × LnGBODSIZE	0.188
	(0.59)
MULT × GINDDIR	–1.625**
	(–2.01)
MULT × GFMLYCHAIR	0.422**
	(2.52)
–1HHI		1.093
		(1.09)
–1HHI × LnGBODSIZE		0.211
		(0.31)
–1HHI × GINDDIR		–3.151*
		(–1.91)
–1HHI × GFMLYCHAIR		0.903**
		(2.57)
Controls	Yes	Yes
Year dummies	Yes	Yes
Observations	726	726
Adjusted R²	.335	.340

Panel B: Financial Statement Comparability.
MULT	0.312***
	(3.53)
MULT × GACCTCOMP	–0.487*
	(–1.73)
–1HHI		0.704***
		(3.60)
–1HHI × GACCTCOMP		–1.210*
		(–1.93)
GACCTCOMP	0.162**	–1.048*
	(2.66)	(–1.68)
Controls	Yes	Yes
Year dummies	Yes	Yes
Observations	726	726
Adjusted R²	.320	.323

Note. This table tests whether the effect of the auditor appointment on the incidence of RPTs varies with board characteristics and financial statement comparability. The dependent variable, LOGRPTFREQ, is defined as the natural log of one plus the number of related-party transactions undertaken by member firms within an FBG in fiscal year t. All variables are defined in the appendix. Each regression includes a separate intercept and the same set of control variables as in Table 3. The t statistics given in parenthesis below each estimate are based on robust standard errors clustering at a group level. RPT = related-party transaction; FBG = family business group.

Denote significance at the 10% level.

Denote significance at the 5% level.

***

Denote significance at the 1% level.

The Role of Financial Statement Comparability

Lower accounting comparability leads to less transparent information environment and increases the costs of market participants to acquire information (Chen et al., 2018; DeFranco et al., 2011). As such, lower accounting comparability could make it more difficult to comprehend accounting numbers in member firms’ financial statements, and in turn prevent auditors from taking timely actions to deter opportunistic intragroup value transfers via RPTs. In this section, we examine whether the adverse impact of multiple-auditor appointments on RPTs is mitigated by higher accounting comparability of group member firms.

We estimate Equation 1 with interactions between GACCTCOMP and each of the two auditor appointment variables. We report the results in Column 1 for MULT and Column 2 for –1HHI in Panel B of Table 6. The coefficients on the interaction variables are negative and statistically significant in both regressions. These results suggest that higher accounting comparability of group member firms can mitigate the impact of multiple-auditor appointments on RPTs.¹⁶

Capturing Intragroup Value Transfers Via Stock Market Reactions

In the previous sections, we capture intragroup value transfers by RPTs. In this section, we capture intragroup value transfers in an indirect way and investigate whether the effect of the auditor appointment strategy on intragroup value transfers still holds. Specifically, we follow G. S. Bae et al. (2008) and use stock price reactions of non-announcing member firms to earnings news released by their group peers to capture the market’s ex ante valuation of capital reallocation within business groups which is referred to as intragroup propping (Friedman et al., 2003; Riyanto & Toolsema, 2008).

Research methodology

Drawing on G. S. Bae et al. (2008), we specify the following regression model:

\begin{matrix} CARN A_{it} = α_{0} + β_{1} CA R_{it} + β_{2} MUL T_{it} / - 1 HH I_{it} \times CA R_{it} + β_{3} MUL T_{it} / - 1 HH I_{it} + \\ β_{4} CASHRIGHT S_{it} + β_{5} FRNHOLDIN G_{it} + β_{6} LnBODSIZ E_{it} + β_{7} INDDI {R_{i}}_{t} + \\ β_{8} FMLYCHAIR + β_{9} SIZ E_{it} + β_{10} LnAUDFE E_{it} + {β_{1}}_{1} PRERE T_{it} + \\ β_{12} ACCTCOM P_{it} + Industry Dummies + Year Dummies + {ε_{i}}_{t} . \end{matrix}

(3)

Similar to G. S. Bae et al. (2008), we use the cumulative abnormal returns of the announcing member firms (CAR) to capture their earnings news, and the cumulative abnormal returns of the non-announcing member firms (CARNA) to measure stock price reactions to the former’s earnings announcements. The relation between CARNA and CAR is expected to capture intragroup propping potential. In particular, we employ the standard market model residuals approach to estimate daily abnormal returns in the event window [–2, 2], where Day 0 is the member firm’s earnings announcement date, and the estimation period spans 200 trading days from Day –210 to Day –11. The local stock market index serves as the market index.¹⁷

We employ the portfolio approach rather than an individual-firm approach to measure the market reactions of non-announcing member firms (CARNA). This avoids the cross-sectional dependence of returns for individual non-announcing member firms.¹⁸CARNA is calculated as the average of the cumulative abnormal returns across all eligible non-announcing member firms around each earnings announcement event.^19,20

We include in Equation 2 the following control variables. CASHRIGHTS is the cash flow rights owned by the controlling family in a member firm. FRNHOLDING is the aggregate equity ownership held by foreign investors in a member firm. LnBODSIZE is the natural log of the number of directors on a member firm’s board. INDDIR is the percentage of independent nonexecutive directors on a member firm’s board. FMLYCHAIR is an indicator variable equal to 1 if a member firm’s board chair is a controlling family member, and 0 otherwise. SIZE is the natural log of a member firm’s market equity value. LnAUDFEE is the natural log of a member firm’s audit fees. PRERET is a member firm’s market-adjusted returns cumulated over 1 year immediately prior to the current fiscal year. ACCTCOMP is the accounting comparability between an announcing member firm and its non-announcing group peer, estimated following DeFranco et al. (2011). Finally, we include industry and year dummies to control for industry and time effects.

Sample and data

Starting with our initial 171 member firms affiliated with 59 FBGs in Hong Kong, we extract earnings announcement dates from the DataStream database. We require that (a) the announcing member firms do not release earnings on the same day as other member firms, (b) non-announcing member firms do not release earnings within 15 days of the earnings announcement dates of their announcing group peers (to avoid information overlap), and (c) stock price data and financial data be available in DataStream. Our final sample consists of 698 earnings announcements made by 107 member firms affiliated with 49 FBGs from 2003 through 2017.

Panel A of Table 7 reports the distribution of sample member firms across auditors. Deloitte accounts for the largest fraction of firm-year observations, followed by PwC, EY, and KPMG. Only 6.16% (43/698) of firm-year observations are audited by non-Big 4 auditors.²¹

Table 7.

Distribution of Sample and Summary Statistics for Market Reaction Tests.

Panel A: Sample Distribution Across Auditors.
Auditor	Member firm-years	Member firms	Family business groups
Deloitte	317	44	26
PwC	148	28	17
EY	110	21	11
KPMG	80	15	9
Non-Big 4	43	16	14
Total	698	124	77

Panel B: Sample Distribution by the Auditor Appointment Strategies.
Group auditor composition	Member firm-years	Member firms	Family business groups
One Big 4	342	61	36
Multiple Big 4	201	39	15
Big 4 and non-Big 4	128	27	9
One non-Big 4	27	4	4
Total	698	131	64

Panel C: Summary Statistics.
Variable	M	SD	Q1	Median	Q3
CARNA	–0.003	0.069	–0.026	–0.006	0.014
CAR	–0.003	0.057	–0.028	0	0.022
MULT	0.471	0.5	0	0	1
–1HHI	–0.756	0.269	–1	–1	–0.44
CASHRIGHTS	0.323	0.191	0.171	0.282	0.454
LnBODSIZE	2.220	0.335	1.946	2.197	2.485
INDDIR	0.382	0.124	0.300	0.364	0.444
FMLYCHAIR	0.506	0.500	0	1	1
SIZE	14.78	1.964	13.407	14.63	16.119
FRNHOLDING	0.117	0.221	0	0	0.130
LnAUDFEE	7.900	1.109	7.102	7.890	8.591
PRERET	0.053	0.543	–0.205	–0.028	0.190
ACCTCOMP	0.328	0.904	0.083	0.164	0.328

Note. This table provides sample distribution and summary statistics for market reaction tests. The sample comprised 698 earnings announcements made by 107 distinct member firms that were affiliated with 49 FBGs in Hong Kong from 2003 through 2017. Panel A reports sample distribution across auditors, Panel B reports sample distribution by auditor appointment strategies of FBGs, and Panel C reports the summary statistics of main variables. All variables are defined in the appendix. FBG = family business group.

Panel B of Table 7 shows the sample distribution across groups’ auditor appointment strategies. In total, 201 earnings announcement events (28.80%) are made by 39 member firms affiliated with 15 FBGs that hire multiple Big 4 auditors. Next, 128 events (18.34%) are made by 27 member firms from nine FBGs that hire both Big 4 and non-Big 4 auditors. Finally, 342 events (49%) are made by 61 member firms from 36 FBGs that hire a single Big 4 auditor, whereas the remaining 27 events (3.86%) are audited by a single non-Big 4 auditor.

Panel C of Table 7 provides descriptive statistics. The mean value of CARNA is –0.003, with a standard deviation of 0.069. The mean value of CAR is –0.003, with a standard deviation of 0.057. About 47.1% of earnings announcements are made by member firms from FBGs that hire multiple auditors (MULT). On average, member firms have 9.749 board members, of which 38.2% are independent directors. Overall, 50.6% of sample member firms have a controlling family member as board chairperson.

Regression results

Table 8 reports the regression results. Column 1 presents the results from regressing CARNA on CAR without control variables included. The coefficient on CAR is significant and positive, consistent with G. S. Bae et al. (2008). This suggests potential intragroup propping.

Table 8.

Regression Results of Market Reaction Tests.

Dependent variable = CARNA	(1)	(2)	(3)	(4)	(5)
CAR	0.143***	0.030	0.030	0.324**	0.325**
	(2.69)	(0.64)	(0.61)	(2.49)	(2.53)
MULT × CAR		0.255**	0.252**
		(2.16)	(2.13)
MULT		0.013**	0.011
		(2.00)	(1.25)
–1HHI × CAR				0.236*	0.241*
				(1.72)	(1.77)
–1HHI				0.011	0.004
				(1.26)	(0.26)
CASHRIGHTS			0.004		–0.001
			(0.26)		(–0.10)
FRNHOLDING			–0.007		–0.007
			(–0.48)		(–0.46)
LnBODSIZE			0.004		0.006
			(0.3)		(0.48)
INDDIR			–0.052		–0.056
			(–1.24)		(–1.36)
FMLYCHAIR			0.009*		0.010*
			(1.70)		(1.77)
SIZE			–0.001		–0.001
			(–0.49)		(–0.30)
LnAUDFEE			–0.004		–0.004
			(–1.06)		(–0.99)
PRERET			–0.006		–0.006
			(–1.54)		(–1.57)
ACCTCOMP			0.006***		0.006***
			(4.02)		(3.75)
Constant	0.000	–0.008	0.048	0.007	0.048
	(–0.03)	(–1.03)	(1.01)	(0.58)	(0.87)
Industry dummies	Yes	Yes	Yes	Yes	Yes
Year dummies	Yes	Yes	Yes	Yes	Yes
Observations	698	698	698	698	698
Adjusted R²	.048	.060	.066	.049	.056

Note. This table presents the results on whether the stock price reactions of non-announcing member firms to earnings news released by their group peers vary with their FBG’s auditor appointment. The sample comprised 698 announcing member firm-year observations that were affiliated with 49 different family business groups in Hong Kong from 2003 through 2017. The dependent variable is CARNA, defined as the average of the 5-day cumulative abnormal returns for member firm i’s non-announcing group peers around its earnings announcement date for fiscal year t. The other variables are defined in the appendix. The t statistics given in parenthesis below each estimate are based on robust standard errors clustering at an announcing member firm level. FBG = family business group.

Denote significance at the 10% level.

Denote significance at the 5% level.

***

Denote significance at the 1% level.

Columns 2 and 4 report the results from regressions with only MULT and –1HHI and their interactions with CAR, whereas Columns 3 and 5 report the results from the full regressions with all variables included. The results are similar with and without the control variables. The coefficient on the interaction variable of interest, MULT×CAR (–1HHI×CAR), is significant and positive in each of these regressions, suggesting that the market perceives member firms’ resources to be more vulnerable to intragroup propping when FBGs hire multiple auditors rather than a single auditor. These results are consistent with those regarding RPTs and support the idea that multiple-auditor appointments weaken the external auditor’s monitoring of intragroup value transfers.²²

Conclusion

This study investigates whether controlling families exploit the auditor appointment strategy to facilitate intragroup value transfers. Using a sample of FBGs in Hong Kong, we find that FBGs that hire multiple auditors tend to make more RPTs than those that hire a single auditor. Our findings suggest that controlling families exploit multiple-auditor appointments as a “divide and conquer” strategy to weaken external auditors’ monitoring against intragroup value transfers. We also find that more independent boards and higher financial accounting comparability of member firms can mitigate the adverse impact of multiple-auditor appointments on intragroup value transfers, but a board chaired by the controlling family member can aggravate the adverse effect. Additional analysis suggests that investors perceive multiple-auditor appointments as impairing audit effectiveness in disciplining intragroup value transfers within FBGs. This study provides new insights regarding the monitoring role of another important auditor choice strategy within FBGs, beyond the choice between Big N versus non-Big N auditors.

Some limitations of our study are noteworthy. First, FBGs could consist of both private and public member firms but our analysis is based on only public firms due to the unavailability of auditor information about private firms. Thus, it might be possible that we misclassify FBGs with multiple auditors as FBGs with a single auditor, although this should work against finding our results. Second, our accounting comparability measure is estimated at the peer group of member firms’ level rather than at the industry level. Because our group member firms could belong to different industries, our results regarding group accounting comparability should be interpreted with caution. Finally, we acknowledge that the generalizability of our findings based on Hong Kong FBGs to other regions might be limited, especially those with formal or informal institutions that are different from those in Hong Kong. Future research can take advantage of cross-country differences in building on this research.

Footnotes

Appendix

Variable Definitions.

Variable	Definition
Auditor appointment variables
MULT	A dummy variable, equal to 1 if multiple independent audit firms audit different member firms within an FBG, and 0 if a single auditor audits all group member firms.
MULTBIG4	A dummy variable, equal to 1 if an FBG hires multiple Big 4 auditors, and 0 otherwise.
BIG4&non-BIG4	A dummy variable, equal to 1 if an FBG hires both Big 4 and non-Big 4 auditors, and 0 otherwise.
–1HHI	The Herfindahl concentration index per FBG based on the number of member clients of each auditor, multiplied by –1. The Herfindahl concentration index = ∑(NCLIENTSi/NFIRMS)², where NCLIENTS is the number of member firms audited by auditor i, and NFIRMS is the number of member firms within an FBG.
RPT frequency variables
RPTFREQ	The number of related-party transactions undertaken by member firms within an FBG in fiscal year t.
LOGRPTFREQ	The natural log of one plus the number of related-party transactions undertaken by member firms within an FBG in fiscal year t.
Abnormal returns variables
CARNA	The average of the 5-day cumulative abnormal returns for member firm i’s non-announcing group peers around its earnings announcement date for fiscal year t.
CAR	The 5-day cumulative abnormal returns for announcing member firm i around its earnings announcement date for fiscal year t.
Other variables
GWEDGE	The median divergence of the controlling family’s control rights from cash flow rights in member firms, where divergence of control rights from cash flow rights is defined as 1 minus cash flow rights divided by control rights.
LnNFIRMS	The log of the number of member firms within an FBG.
LnNSECTORS	The log of the number of industries (two-digit SIC Code) in which an FBG’s member firms operate.
LnNBLOCKS	The natural log of one plus the number of member firms with at least one blockholder other than their controlling family holding no less than 5% cash flow rights within an FBG.
LnBODSIZE	The natural log of the number of board directors of member firm i in fiscal year t.
LnGBODSIZE	The natural log of the median number of board directors of member firms within an FBG in fiscal year t.
INDDIR	The fraction of independent directors of member firm i in fiscal year t.
GINDDIR	The median of the fraction of independent directors of member firms within an FBG in fiscal year t.
GFMLYCHAIR	The fraction of member firms within an FBG that have a controlling family member serving as board chairperson in fiscal year t.
FMLYCHAIR	A dummy variable, equal to 1 if member firm i has a controlling family member serving as board chairperson in fiscal year t, and 0 otherwise.
SIZE	The natural log of member firm i’s market capitalization at the end of fiscal year t – 1.
GSIZE	The natural log of median market capitalization of member firms within an FBG at the end of fiscal year t – 1.
GLEV	The median of ratios of total liabilities to total assets for member firms within an FBG at the end of fiscal year t – 1.
GMB	The median of market-to-book ratio for member firms within an FBG at the end of fiscal year t – 1.
GROA	The median of returns on assets for member firms within an FBG at the end of fiscal year t – 1.
LnAUDFEE	The natural log of auditor fees (in millions) paid by member firm i in fiscal year t – 1.
LnGAUDFEE	The natural log of the median auditor fees (in millions) paid by member firms within an FBG in fiscal year t – 1.
GACCTCOMP	The median of the member firms’ accounting comparability within an FBG in fiscal year t – 1, where accounting comparability is estimated following DeFranco et al. (2011).
DISFSALES	The standard deviation of the member firm’s ratio of foreign sales to total sales within an FBG in year t – 1.
CASHRIGHTS	Cash flow rights of member firm i held by its controlling family in year t.
FRNHOLDING	Equity ownership held by foreign investors in year t.
PRERET	Member firm i’s market-adjusted returns (difference between realized returns and market returns) cumulated over 1 year immediately prior to the fiscal year t.

Acknowledgements

The authors gratefully acknowledge insightful suggestions from the editor (Bharat Sarath), associate editor (Linda Myers), and anonymous reviewers who have helped us to improve the earlier version of this manuscript. The authors appreciate helpful comments from the seminar participants at Hong Kong Baptist University, Zhongnan University of Economics and Law, University of Macau, and Hang Seng University of Hong Kong. Haw acknowledges that parts of his work were conducted during his visit to Hang Seng University of Hong Kong. The authors alone are responsible for all limitations and errors that may relate to the paper. All coauthors have made equal contribution to the paper.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Tan thanks for the financial support from Guangdong Basic and Applied Basic Research Foundation (Project No.: 2019A1515011990). Wang acknowledges the financial support from Hong Kong Research Grants Council (Ref. 12502714).

ORCID iD

In-Mu Haw

Notes

References

Bae

G. S.

Cheon

Y. S.

Kang

J.-K.

(2008). Intragroup propping: Evidence from the stock-price effects of earnings announcements by Korean business groups. Review of Financial Studies, 21, 2015-2060.

Bae

Kang

J.-K.

Kim

(2002). Tunneling or value added? Evidence from mergers by Korean business groups. Journal of Finance, 57, 2695-2740.

Baek

Kang

J.-K.

Lee

(2006). Business groups and tunneling: Evidence from private securities offerings by Korean Chaebols. Journal of Finance, 61, 2415-2449.

Becker

DeFond

Jiambalvo

Subramanyam

(1998). The effect of audit quality on earnings management. Contemporary Accounting Research, 15, 1-24.

Bedard

Johnstone

(2004). Earnings manipulation risk, corporate governance risk, and auditors’ planning and pricing decisions. The Accounting Review, 79, 277-304.

Bell

Doogar

Solomon

(2008). Audit labor usage and fees under business risk auditing. Journal of Accounting Research, 46, 729-760.

Bell

Landsman

Shackelford

(2001). Auditors’ perceived business risk and audit fees: Analysis and evidence. Journal of Accounting Research, 39, 35-43.

Bennouri

Nekhili

Touron

(2015). Does auditor reputation “discourage” related-party transactions? The French case. Auditing: A Journal of Practice and Theory, 34, 1-32.

Bertrand

Mehta

Mullainathan

(2002). Ferreting out tunneling: An application to Indian business groups. The Quarterly Journal of Economics, 117, 121-148.

10.

Boone

Khurana

Raman

(2011). Audit market concentration and auditor tolerance for earnings management. Contemporary Accounting Research, 20, 1-33.

11.

Byrd

Hickman

(1992). Do outside directors monitor managers? Evidence from tender offer bids. Journal of Financial Economics, 32, 195-221.

12.

Carney

(2008). Asian business groups: Context, governance and performance. Chandos Publishing.

13.

Chang

Hong

(2000). Economic performance of group-affiliated companies in Korea: Intragroup resource sharing and internal business transactions. The Academy of Management Journal, 43, 429-448.

14.

Chen

Collins

Kravet

Mergenthaler

(2018). Financial statement comparability and the efficiency of acquisition decisions. Contemporary Accounting Research, 35, 164-202.

15.

Cheung

Y.-L.

Haw

Tan

Wang

(2014). Board structure and intragroup propping: Evidence from family business groups in Hong Kong. Financial Management, 43, 569-601.

16.

Cheung

Y.-L.

Rau

Stouraitis

(2006). Tunneling, propping, and expropriation: Evidence from connected party transactions in Hong Kong. Journal of Financial Economics, 82, 343-386.

17.

Choi

J. H.

Kim

J. B.

Zang

(2010). Do abnormally high audit fees impair audit quality? Auditing: A Journal of Practice and Theory, 29, 115-140.

18.

Choi

Wong

(2007). Auditors’ governance functions and legal environments: An international investigation. Contemporary Accounting Research, 2, 41-36.

19.

Claessens

Djankov

Fan

Lang

(2002). Disentangling the incentive and entrenchment effects of large shareholdings. Journal of Finance, 57, 2741-2771.

20.

Claessens

Djankov

Lang

(2000). The separation of ownership and control in East Asian corporations. Journal of Financial Economics, 58, 81-112.

21.

Dahya

Dimitrov

McConnell

(2008). Dominant shareholders, corporate boards, and corporate value: A cross-country analysis. Journal of Financial Economics, 87, 73-100.

22.

DeFond

Francis

Wong

(2000). Auditor industry specialization and market segmentation: Evidence from Hong Kong. Auditing: A Journal of Practice and Theory, 19, 49-66.

23.

DeFranco

Kothari

Verdi

(2011). The benefits of financial statement comparability. Journal of Accounting Research, 49(4), 895-931.

24.

Djankov

La Porta

Lopez-de-Silanes

Shleifer

(2008). The law and economics of self-dealing. Journal of Financial Economics, 88, 430-465.

25.

Fama

E. F.

Jensen

(1983). Separation of ownership and control. Journal of Law and Economics, 26, 301-325.

26.

Fan

Wong

(2002). Corporate ownership structure and the informativeness of accounting earnings in East Asia. Journal of Accounting and Economics, 33, 401-425.

27.

Fan

Wong

(2005). Do external auditors perform a corporate governance role in emerging markets? Evidence from East Asia. Journal of Accounting Research, 43, 35-72.

28.

Fang

Lobo

Zhang

Zhao

(2018). Auditing related party transactions: Evidence from audit opinions and restatements. Auditing: A Journal of Practice and Theory, 37, 73-106.

29.

Fang

Pittman

Zhang

Zhao

(2017). Auditor choice and its implications for group affiliated firms. Contemporary Accounting Research, 34, 39-82.

30.

Fisman

Wang

(2010). Trading favors within Chinese business groups. American Economic Review: Papers & Proceedings, 100, 429-433.

31.

Francis

Anderson

Stokes

(1999). City markets as a unit of analysis in audit research and the re-examination of Big 6 market shares. Abacus, 35, 185-206.

32.

Francis

Pinnuck

Watanabe

(2014). Auditor style and financial statement comparability. The Accounting Review, 89, 605-633.

33.

Francis

Michas

Seavey

(2013). Does audit market concentration harm the quality of audited earnings? Evidence from audit markets in 42 countries. Contemporary Accounting Research, 30, 325-355.

34.

Francis

Wang

(2008). The joint effect of investor protection and Big 4 audits on earnings quality around the world. Contemporary Accounting Research, 25, 157-191.

35.

Friedman

Johnson

Mitton

(2003). Propping and tunneling. Journal of Comparative Economics, 31, 732-750.

36.

Gillette

Noe

Rebello

(2003). Corporate board composition, protocols, and voting behavior: Experimental evidence. Journal of Finance, 58, 1997-2032.

37.

Gopalan

Nanda

Seru

(2014). Internal capital market and dividend policies: Evidence from business groups. Review of Financial Studies, 27, 1102-1142.

38.

Guedhami

Pittman

Saffar

(2014). Auditor choice in politically connected firms. Journal of Accounting Research, 52, 107-162.

39.

Habib

Jiang

Zhou

(2015). Related-party transactions and audit fees: Evidence from China. Journal of International Accounting Research, 14, 59-83.

40.

Harris

Raviv

(2008). A theory of board control and size. Review of Financial Studies, 21, 1797-1832.

41.

Haw

Hwang

(2004). Ultimate ownership, income management, and legal and extra-legal institutions. Journal of Accounting Research, 42, 423-462.

42.

Haw

(2011). Corporate governance and earnings management by classification shifting. Contemporary Accounting Research, 28, 517-553.

43.

Jian

Wong

(2010). Propping through related party transactions. Review of Accounting Studies, 15, 70-105.

44.

Joe

Vandervelde

(2007). Do auditor-provided nonaudit services improve audit effectiveness? Contemporary Accounting Research, 24, 467-487.

45.

Johnson

La Porta

Lopez-De-Silanes

Shleifer

(2000). Tunneling. American Economic Review, 90, 22-27.

46.

Kim

(2006). Ownership structure, business group affiliation, listing status, and earnings management: Evidence from Korea. Contemporary Accounting Research, 23, 427-464.

47.

Kim

Pevzner

Xin

(2019). Foreign institutional ownership and auditor choice: Evidence from worldwide institutional ownership. Journal of International Business Studies, 50, 83-110.

48.

King

(2002). An experimental investigation of self-serving biases in an auditing trust game: The effect of group affiliation. The Accounting Review, 77, 265-284.

49.

Kohlbeck

Mayhew

(2017). Are related party transactions red flags? Contemporary Accounting Research, 34, 900-928.

50.

Wong

Firth

(2010). Tax, financial reporting, and tunneling incentives for income shifting: An empirical analysis of the transfer pricing behavior of Chinese-listed companies. Journal of the American Taxation Association, 32, 1-26.

51.

Morck

Yeung

(2003). Agency problems in large family business groups. Entrepreneurship Theory and Practice, 27, 367-382.

52.

Morck

Wolfenzon

Yeung

(2005). Corporate governance, economic entrenchment, and growth. Journal of Economic Literature, 43, 655-720.

53.

Nelson

Elliott

Tarpley

(2002). Evidence from auditors about managers’ and auditors’ earnings management decisions. The Accounting Review, 77, 175-202.

54.

Posner

Spier

Vermeule

(2010). Divide and conquer. Journal of Legal Analysis, 2, 417-471.

55.

Prencipe

Bar-Yosef

(2011). Corporate governance and earnings management in family-controlled companies. Journal of Accounting, Auditing and Finance, 26, 199-227.

56.

Raheja

C. G.

(2005). Determinants of board size and composition: A theory of corporate boards. Journal of Financial and Quantitative Analysis, 40, 283-306.

57.

Riyanto

Toolsema

(2008). Tunneling and propping: A justification for pyramidal ownership. Journal of Banking and Finance, 32, 2178-2187.

58.

Shin

H.-H.

Park

(1999). Financing constraints and internal capital markets: Evidence from Korean Chaebols. Journal of Corporate Finance, 5, 169-191.

59.

Tajfel

Turner

(1986). The social identity theory of intergroup behavior. In Worchel

Austin

W. G.

(Eds.), Psychology of intergroup relations (2nd ed., pp. 7-24). Nelson-Hall Publishers.

60.

Wooldridge

J. M.

(2016). Introductory econometrics: A modern approach. South-Western College Publishing.