Chair–CEO trust and firm performance

Abstract

This study investigates whether trust between a corporation’s board chair and the CEO affects firm performance. After using a unique survey data set of regional trustworthiness from China to measure this trust, we find a positive relationship between trust and the performance of Chinese companies from 2000 to 2016. Additional test results suggest that the relationship is causal. Further results show that the positive trust-performance effect is more evident for firms with greater advisory needs and boards that can deliver high-quality advice. Finally, we find supporting evidence to our conjecture that the Chair–CEO trust increases firm value by improving the board advisory results, including value-adding decisions of R&D and merger and acquisition.

JEL Classification: G32, G34, G41

Keywords

Advisory role firm performance merger and acquisition R&D social trust

1. Introduction

Trust determines the tendency of people to cooperate (Coleman, 1990; Putnam, 1993). It is regarded as essential to social capital (Guiso et al., 2004) to the extent that economists have long been interested in the payoffs of trust. This article tests whether trust between a CEO and the chair of a board of directors, the two most prominent decision-makers in a firm, drives firm performance. Our study is motivated by the “friendly” board theory of Adams and Ferreira (2007) wherein a management-friendly board encourages its CEO to share information with it. Based on this reasoning, we conjecture that a friendly board is likely to form in a company when it trusts the CEO highly. This leads to a strong advisory board process, especially regarding information sharing between the CEO and the board. As a result, trust promotes value-adding strategic decisions such as research and development (R&D), and merger and acquisition (MA).

The primary purpose of this study is to investigate whether trust between the CEO and the board chair affects the advisory outcome of a firm. We measure the trust level between them according to how the regional trust their provinces of origin are ranked. To do so, we use the regional trust index created by Zhang and Ke (2002) but which is unavailable in other countries. In this study, we use the setting of China because Zhang and Ke’s regional trust data allow us to respectively determine how a board chair perceives a CEO’s trustworthiness (measured as “Chair–CEO trust”) and how a CEO perceives a board chair’s trustworthiness (“CEO–Chair trust”). Zhang and Ke’s trust index is based on the answers to questions about the regional social trust of 5000 managers and entrepreneurs from 31 provinces in China. Owing to the provinces’ various proximities to each other, and their ethnic, dialectical, regionally cultural, and historical diversity in China, these surveyees’ perceptions of regional trustworthiness result in large variations in their trust level.

Another benefit of using the setting of China is that Chinese companies have a dual governance structure that comprises a management board and a supervisory board. Because these two boards are independent of each other, the responsibilities of advising and monitoring can be better separated than those with the single board system that occurs in some other countries (e.g. the United States). Therefore, the goal of the management committee is to advise the CEO, while the supervisory board supposedly monitors the board of directors and the management team. However, the literature suggests that supervisory boards play no significant monitoring role in China (Farag and Mallin, 2017)¹ and thus, in China, do not improve firm value (Bai et al., 2004; Li and Naughton, 2007). Hence, our results focusing on the advisory effect of the management board chair can be feasibly isolated from the (weak) monitoring effect of the supervisory board. The chair of the management committee and the CEO are regarded as the two most important decision-makers in Chinese companies. We use the trust between them to proxy for the advisory relationship between the board and the management.

We first evaluate whether the trust between the CEO and the chair of the management board is a driving force of company performance. Our baseline regression results suggest positive associations of both directions of trust on firm performance, with Chair–CEO trust being more significant than CEO–Chair trust. The result is not only statistically but also economically significant. An increase of one standard deviation in Chair–CEO trust is associated with an increase in return on equity (ROE) (net profit margin (NPM) and return on asset (ROA)) of 1.61% (0.55% and 0.41%). We next examine whether the trust is more influential for firms with (1) greater advisory needs, for example, diversified firms (Coles et al., 2008; Schmidt, 2015) and (2) boards having a higher proportion of busy boards, which, according to Field et al. (2013), can provide high-quality advice. Our results provide supporting evidence to both, indicating that trust is a driving force of firm performance through its impact on the advisory board process.

Our empirical specifications include year-fixed effects and industry-fixed effects, with standard errors clustered on both firm and year to correct for cross-sectional dependence and serial correlation. To address the concern of simultaneity, we regress firm performance measures on the one-period lag values of all explanatory accounting variables. To mitigate other concerns of endogeneity, we adopt the two-stage least squares method (2SLS) with two instrumental variables constructed for Chair–CEO trust: (1) the geographical distance between birthplaces of the CEO and the chair within a company and (2) an indicator to identify whether they are from the same or neighboring provinces. Our trust-performance finding holds after adopting the instrumental variable (IV) approach. In addition, we conduct a placebo test by randomly assigning a Chair–CEO pair to a sample firm-year. Since the allocations of the CEO–Chair pair are random, we know that the null of the relationship between trust and firm performance is zero. We find that the mean of the 1000 estimates of the effect of the pseudo Chair–CEO trust on firm performance is zero. Values of all the 1000 coefficients of simulation results are smaller than the estimated coefficient using the actual sample, that is, a zero in 1000 chance of observing the actual trust effect. The placebo test results further confirm that our finding is not caused by reverse causality that well-performing firms seek to appoint trustworthy CEOs and chairs.

Results on additional robustness tests also support this causal interpretation. First, we exploit CEO turnovers by including the previous Chair–CEO turnover trust in the regression model as an additional explanatory variable to alleviate concerns that our results might be driven by omitted variables that can, for example, induce self-selection bias caused by the missing birthplace demographic information of a chair or CEO. Second, we show that our finding is not driven by economic growth and educational resources of birthplaces of the CEOs and the board chairs. Third, we demonstrate that our conclusion is unaffected after excluding observations where a firm’s CEO and management board chair are from the same province. Fourth, we also confirm that our results are not driven by the political and educational backgrounds of the chair and the CEO, or their local connections, including their alumni relations. Finally, we exploit the event of CEO turnovers to see whether our result is driven by pre-turnover Chair–CEO trust, an omitted variable. The result does not support this concern and we find that only the current trust relationship impacts firm performance.

To explore the mechanism for the found relationship, we study two strategic business decisions of R&D and MA, in which the firm is most likely to benefit from a strong advisory process. R&D is a dominant source of competitive advantage, especially for technology companies (Lin et al., 2006). Investment in R&D requires expert opinions, usually relates to the advisory opinions of the board of directors (Chen, 2015). Similarly, decisions about MA must be approved by the management board; MA decisions directly affect the company’s performance (Masulis and Mobbs, 2011). We find that a higher degree of trust between the company’s chair and its CEO will enhance R&D outcomes, especially for technology companies. We also observe that the market reacts more strongly to MA announcements by companies with a high level of Chair–CEO trust.

This study ties into several strands of the literature. First, it complements the literature of social norms and trust with research investigating the relationship between trust and firm value. We focus on social trust at the individual level, which differs from recent research on the trustworthiness of the region where the corporate headquarters is located. Studies in this stream of the literature find that firms located in regions of high social trust are associated with lower likelihoods of corporate misconduct (Dong et al., 2016; Qiu et al., 2019), tax avoidance (Xia et al., 2017), corruption (LaPorta et al., 1997), earning misstatement (Garrett et al., 2014), and bad news, hoarding behaviors of managers (Li et al., 2017).²

Second, we provide new empirical evidence regarding the importance of the advisory role of a board and the benefits of having a management-friendly board (Adams and Ferreira, 2007; Baldenius et al., 2014; Harris and Raviv, 2008; Kang et al., 2018; Schmidt, 2015). Finally, our study reveals that such a relationship between the CEO and the board chair can impact firm value; research on the importance of this relationship is inconclusive.³ Personal relationships studied in these papers are identified as overlaps of personal experience (e.g. the CEO and the chair are alumni) or social contacts (e.g. they joined the same country clubs). These identification methods are usually constrained by the availability and reliability of data on the social networks of the CEO and the board. Moreover, measures to proxy for this relationship are often monotonic, being merely an indicator to show whether there is an overlap in their social networks. This current study is less constrained regarding data availability, and the trust data in use show heterogeneity across regions.

2. Sample and descriptive statistics

2.1. Sample and data sources

Our sample consists of all Chinese A-share listed companies from 2000 to 2016. From the China Securities Market and Accounting Research (CSMAR) database, we obtain information about sample companies (financial and market data) and their CEOs and board chairs (e.g. age, birthplace, tenure, education/political background, and overseas working experience). We exclude (1) firms that do not disclose their chair’s or CEO’s birthplace (73.66%), in which case our trust measure cannot be constructed; (2) firm-year observations with missing values for dependent or explanatory variables (8.88%); and (3) financial firms (0.34%) as the industry is highly regulated and their objectives can be different from other industries.

Our final sample consists of 2812 firm-year observations with CEOs or management board chairs from all 28 provinces in China during the period of 2000 to 2016. All continuous variables are winsorized at the 1% and 99% levels. Figure 1(a) illustrates the distribution of our sample by industry, revealing that information technology and machinery are the two largest industries in the sample.

Figure 1.

Industry distribution and yearly distributions of Chair–CEO trust and CEO–chair trust: (a) Industry distribution. (b) Yearly distributions of Chair–CEO trust and CEO–chair trust.

2.2. Measuring trust between the board chair and the CEO

We measure Chair–CEO trust using regional trust data from Zhang and Ke (2002).⁴ Their trust data are adopted in some of latest trust research on the Chinese market, including Jiang et al. (2020), Li et al. (2017), and Chen et al. (2017). Our proxy for Chair–CEO trust, denoted by $T r u s t_{c h a i r}$ , is calculated as the natural logarithm of cross-provincial-level trustworthiness based on the birthplaces of the board chair and the CEO. For example, Appendix 1 shows that people from Shanghai evaluate the trustworthiness level of people from Beijing as 7.4 (row 9, column 2). If, in our sample, a company has a Shanghai-born board chair and a Beijing-born CEO, the $T r u s t_{c h a i r}$ value is 2 (i.e. the natural logarithm of the Beijing–Shanghai trustworthy level 7.4). Likewise, for a company with a Beijing-born board chair and a Shanghai-born CEO, its $T r u s t_{c h a i r}$ value is 3.17, calculated as the natural logarithm of the Shanghai–Beijing trustworthy level of 23.9 (row 2, column 10 of Appendix 1). This example illustrates that our measure of trust is bi-directional, which can be asymmetric in value when a board chair and a CEO come from different regions. It is worth noting that it is not uncommon that one can more highly trust those from different regions than people from the same region. For example, people from Hebei trust those from Beijing (i.e. a trustworthy rating of 27.7) more than people also from Hebei (a rating of 17). Distributions of chair–CEO trust and CEO–chair trust over time (see Figure 1, Panel B) confirm this finding, as in some years the average chair–CEO trust level is higher (or lower) than the average CEO–chair trust.

Panel A of Table 1 reports the summary statistics for our sample firms. Both $T r u s t_{c h a i r}$ and $T r u s t_{C E O}$ exhibit variations across the sample. Notably, the correlation of $T r u s t_{c h a i r}$ and $T r u s t_{C E O}$ reveals that although the bi-directional trust levels are asymmetric, the two measures are highly correlated (0.95). This high correlation suggests that mutual trust is likely to be established when a board chair highly trusts the CEO and vice versa.

Table 1.

Summary statistics and Pearson pairwise correlations.

Panel A: Summary statistics on the full sample during the sample period of 2000–2016.
	Obs.	M	SD	Min	Max
TrustChair	2812	1.778	1.214	0.182	4.366
TrustCEO	2676	1.812	1.211	0.182	4.366
Trustcompany	2812	84.175	58.817	4.100	218.900
ROE	2812	0.084	0.138	−4.666	0.525
NPM	2812	0.052	0.057	−0.284	0.331
ROA	2812	0.053	0.050	−0.278	0.296
Size	2812	15.551	1.114	12.645	21.217
Firmage	2812	1.659	0.646	0.000	2.833
FIN_lev	2812	0.406	0.211	0.025	1.122
BM	2812	0.847	0.932	0.032	11.005
Growth_rate	2812	0.335	2.710	−0.915	103.812
Volatility	2812	0.580	0.238	0.135	3.043
GDP_Growth	2812	0.110	0.055	−0.223	0.370
Stateown	2812	0.373	0.149	0.081	0.765
Largestshare	2812	0.267	0.443	0.000	1.000
Boardsize	2812	2.152	0.213	1.386	2.890
Duality	2812	0.606	0.489	0	1
Managerialownership	2812	0.026	0.079	0.144	1.168

Panel B: Pearson correlations (*p < 0.05).
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
(1) TrustChair	1
(2) TrustChair	0.9468*	1
(3) Trustcompany	0.1917*	0.1946*	1
(4) ROE	0.0258	0.0153	0.0376*	1
(5) NPM	0.0051	−0.0048	0.0548*	0.7315*	1
(6) ROA	0.0155	0.0146	0.0185	0.7307*	0.9566*	1
(7) Size	−0.0117	0.000300	0.2145*	0.2723*	0.2555*	0.2584*	1
(8) FIN_lev	0.0533*	0.0835*	−0.0077	−0.1190*	−0.3680*	−0.2262*	0.1818*	1
(9) Growth_rate	0.0144	0.0137	−0.0005	0.0922*	0.0832*	0.0732*	0.0384*	0.0406*	1
(10) Firmage	0.0174	0.0271	−0.0533*	−0.1308*	−0.2612*	−0.1426*	0.2621*	0.3765*	0.0259
(11) BM	0.0642*	0.0659*	0.1057*	−0.1290*	−0.3105*	−0.2508*	0.1028*	0.5376*	−0.0094
(12) Volatility	−0.0589*	−0.0721*	0.0128	0.0291	0.0521*	0.0292	0.0986*	−0.0763*	0.0356*
(13) Boardsize	0.1276*	0.1559*	−0.0838*	0.0720*	0.0281	0.0649*	0.1072*	0.1530*	−0.0227
(14) Largeshare	0.0666*	0.0653*	0.0542*	0.1449*	0.0981*	0.1069*	0.1679*	0.0775*	0.0444*
(15) Stateown	0.0886*	0.1097*	−0.0387*	0.0042	−0.0816*	−0.0309	0.1421*	0.2779*	0.1680*
(16) GDP_Growth	−0.0272	−0.0276	−0.2581*	0.1107*	0.1298*	0.1380*	−0.2457*	−0.0096	0.0218
(17) Duality	−0.5145*	−0.5418*	0.1103*	0.0436*	0.1025*	0.0543*	−0.0712*	−0.2152*	0.0071
(18) Managerialownership	−0.0559*	−0.0594*	0.0383*	−0.0596*	0.0180	0.0703*	−0.0080	−0.0129	−0.0080
	(10)	(11)	(12)	(13)	(14)	(15)	(16)	(17)	(18)
(10) Firmage	1
(11) BM	0.2406*	1
(12) Volatility	−0.0336*	−0.2318*	1
(13) Boardsize	0.0073	0.1328*	−0.1349*	1
(14) Largeshare	−0.0575*	0.1474*	−0.1235*	0.0776*	1
(15) Stateown	0.1723*	0.2797*	−0.2028*	0.3390*	0.2386*	1
(16) GDP_Growth	−0.2281*	−0.0183	−0.1073*	0.1271*	0.0802*	0.1680*	1
(17) Duality	−0.1610*	−0.2045*	0.1471*	−0.2440*	−0.1311*	−0.3606*	−0.1073*	1
(18) Managerialownership,t	−0.0144	−0.1037*	0.0709*	−0.0814*	−0.0187	−0.0894*	−0.0397*	0.1025*	1

Panel A of the table presents the descriptive statistics on the main variables in this study. The sample period is from 2000 to 2016. All variables are defined in Appendix 3. Panel B presents the Pearson correlations of the variables.

2.3. Measuring firm-level trust as a control

Several studies use the same trust data but for different purposes (Chen et al., 2016; Li et al., 2017; Wu et al., 2014), mainly to examine whether social trust at a firm level can curb unethical management behavior. To control for the impact of firm-level trust on firm performance, we include $T r u s t_{c o m p a n y}$ in our regression model. Following Wu et al. (2014) and Li et al. (2017), we determine $T r u s t_{c o m p a n y}$ using the natural logarithm of the regional trust score based on the company headquarter’s location. For example, if the company is located in Beijing, which has a trust score of 169, the relevant value of $T r u s t_{c o m p a n y}$ is 5.13 (i.e. calculated as the natural logarithm of the trust score: ln(169) = 5.13). Appendix 2.1 presents the trust scores for each province. They are calculated by the weighted average of the rankings provided by managers (Zhang and Ke, 2002). For example, Beijing is ranked number one by 16.6% of the responding managers, number two by 11.3% of them, number three by 8.3%, number four by 5.5%, and number five by 4.9% of them. Therefore, the trust score of Beijing is 169% (calculated as 1(16.6% × 5) + (11.3% × 4) + (8.3% × 3) + (5.5% × 2) + (4.9% × 1)). Panel B of Table 1 shows a moderate correlation of 0.19 between $T r u s t_{c h a i r}$ and $T r u s t_{c o m p a n y}$ , indicating that the trust level based on the firm’s geographical location (i.e. regional trust) differs substantially from the personal trust.

2.4. Measuring firm performance as dependent variables

We measure firm performance using three variables that are well established in the literature: ROE, NPM, and ROA. Firms with higher ROA, ROE, or NPM are recognized as better performing firms. ROE represents net profit divided by the average balance of shareholder’s equity (Baliga et al., 1996). NPM is measured by the net profits divided by the average balance of total assets (Rechner and Dalton, 1991). ROA is the sum of total profits and financial expenses divided by the average balance of total assets; it is recorded for each fiscal year (Daily and Dalton, 1994; Quigley and Hambrick, 2011). These performance measures have been used widely in the literature (e.g. Beck et al., 2018; Jia and Bradbury, 2020; Khan et al., 2014; Matolcsy and Wright, 2011; Park and Lee, 2020). The mean values of these, our primary dependent variables, are 0.084 (ROE), 0.052 (NPM), and 0.053 (ROA) (see Panel A, Table 1). Their mean and standard deviations are similar to other large-sample studies of the Chinese market (e.g. Li et al., 2017; Liu et al., 2016). The pairwise correlations of our three measurements of firm performance are high, with coefficients ranging from 0.73 to 0.96 (Panel B, Table 1).

3. Does trust between the management board chair and the CEO drive firm performance?

3.1. The effect of Chair–CEO trust on firm performance

Considering the unique characteristics of the Chinese setting, where the chair plays a crucial role on the board, we assume that the chair’s trust in the CEO (i.e. Chair–CEO trust) represents the board’s trust in the CEO. Regression Model (1) is used to test the relationship between Chair–CEO trust ( $T r u s t_{c h a i r}$ ) and firm performance

F P_{i, t + 1} = β_{0} + β_{1} T r u s t_{c h a i r, i t} + γ C o n t r o l s_{i t} + I n d u s t r y + Y e a r + ε_{i, t + 1}

Model (1)

where the dependent variable $F P_{i, t + 1}$ proxies for the firm performance of company i that is respectively measured by ROE, NPM, and ROA in year t + 1. Our primary variable of interest is $T r u s t_{c h a i r}$ that measures the trust of the board chair in the CEO of a company at year t. To mitigate the concern of simultaneity, we measure all other explanatory variables in year t, representing a 1-year lag from the dependent variable. The control variables include factors derived from the literature related to predicting firm performance. First, these include firm-level controls: firm size (Size), financial leverage (Finlev), book-to-market ratio (BM), firm listing age (Firm_age), the growth rate in operating income in the current year from the previous year (Growth_rate), and stock volatility (Volatility).⁵ Second, these consist of corporate governance and ownership controls: board size (Boardsize), state ownership (Stateown), ownership concentration (Largest), and CEO duality (Duality).⁶ Third, regional economic growth rate ( $G D P_{g r o w t h}$ ) and the growth rate of gross domestic product (GDP) per capita for each province where a company is headquartered are included as an additional control because the regional economic development level can influence local firm performance (Bonin and Huang, 2002). Definitions of variables are detailed in Appendix 3. In the regressions, we also control for industry-fixed effects and year-fixed effects. Standard errors are clustered at both firm and year levels.

Panel A of Table 2 presents our baseline regression results regarding the relationship between Chair–CEO trust and firm performance. We find positive and significant relationships between $T r u s t_{c h a i r}$ and all three measures of firm performance (column 1). To mitigate the concern that our results are driven by the impact of social trust at the firm level, we add $T r u s t_{c o m p a n y}$ into the regressions. In addition, we adopt the 2SLS method to mitigate the concern of endogeneity. To implement, we create two instrument variables for the trust measure: (1) the geographical distance between the birthplaces of CEO and the chair and (2) an indicator where a value of one is assigned to observations when the CEO and the chair are from same or neighboring provinces. We expect that people living near each other would interact more than those living comparatively further apart. These interactions help create common grounds, such as shared culture and trust.

Table 2.

The relationship between Chair–CEO trust and firm performance.

Panel A: The Chair–CEO trust and firm performance.
Dependent variables:	ROEt+1			NPMt+1		ROAt+1
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
	OLS	OLS	IV-2SLS	OLS	IV-2SLS	OLS	IV-2SLS
TrustChair,t	0.0107**	0.0133**	0.0065*	0.0045***	0.0038**	0.0034**	0.0030*
	(2.25)	(2.33)	(1.82)	(2.91)	(2.13)	(2.37)	(1.79)
Trustcompany,t		−0.0002	−0.0001*	−0.0000	−0.0000	−0.0001*	−0.0000*
		(−1.62)	(−1.69)	(−1.33)	(−1.33)	(−1.68)	(−1.75)
Sizet	0.0597***	0.0621***	0.0564***	0.0301***	0.0293***	0.0249***	0.0244***
	(11.98)	(10.36)	(26.85)	(18.75)	(27.76)	(16.70)	(24.96)
FIN_levt	−0.0938	−0.0959	−0.0387***	−0.0921***	−0.0850***	−0.0511***	−0.0460***
	(−1.62)	(−1.63)	(−3.69)	(−8.31)	(−16.12)	(−4.82)	(−9.41)
Growth_ratet	0.0025**	0.0024**	0.0023***	0.0009**	0.0009***	0.0006	0.0006**
	(2.57)	(2.56)	(3.75)	(2.18)	(2.91)	(1.60)	(1.98)
Firmaget	−0.0001	−0.0008	−0.0051	−0.0025	−0.0034**	0.0027	0.0017
	(−0.02)	(−0.12)	(−1.58)	(−0.89)	(−2.10)	(0.91)	(1.15)
BMt	−0.0192**	−0.0178**	−0.0225***	−0.0110***	−0.0114***	−0.0113***	−0.0114***
	(−2.44)	(−2.16)	(−9.33)	(−3.93)	(−9.43)	(−4.36)	(−10.13)
Volatilityt	0.0051	0.0047	0.0063	0.0045	0.0043	0.0002	0.0004
	(0.50)	(0.47)	(0.72)	(0.86)	(0.97)	(0.03)	(0.10)
Boardsizet	0.0326	0.0280	0.0131	0.0045	0.0037	0.0059	0.0052
	(1.44)	(1.35)	(1.50)	(0.55)	(0.85)	(0.79)	(1.28)
Dualityt	0.0291*	0.0342*	0.0175***	0.0099***	0.0102***	0.0073**	0.0080***
	(1.80)	(1.84)	(2.68)	(2.70)	(3.10)	(2.12)	(2.63)
Largestsharet	0.0737***	0.0760***	0.0581***	0.0148	0.0118*	0.0132	0.0097*
	(2.67)	(2.70)	(4.80)	(1.26)	(1.94)	(1.23)	(1.71)
Stateownt	−0.0132	−0.0128	−0.0194***	−0.0114***	−0.0108***	−0.0120***	−0.0112***
	(−1.24)	(−1.17)	(−3.86)	(−2.86)	(−4.29)	(−3.11)	(−4.78)
GDPgrowth,t	0.2123**	0.1380*	0.1612***	0.0484	0.0658**	0.0322	0.0487*
	(2.31)	(1.79)	(3.00)	(1.24)	(2.43)	(0.89)	(1.94)
Managerialownership,t	0.0499	0.0499	0.0525**	0.0058	0.0078	0.0123	0.0144
	(0.53)	(0.55)	(2.44)	(0.14)	(0.72)	(0.30)	(1.44)
Industry dummies	YES	YES	YES	YES	YES	YES	YES
Year dummies	YES	YES	YES	YES	YES	YES	YES
_Cons	−0.9761***	−0.9847***	−0.7536***	−0.3633***	−0.3483***	−0.3045***	−0.2931***
	(−7.98)	(−7.77)	(−19.11)	(−11.22)	(−17.55)	(−10.17)	(−15.94)
First stage—Natural logarithm of distance between the birthplaces of CEO and the chair			−0.2662***(−29.02)		−0.2662***(−29.02)		−0.2662***(−29.02)
Obs.	2820	2812	2758	2812	2758	2813	2759
Adjusted R2	0.224	0.227	0.323	0.400	0.403	0.321	0.331

Panel B: The CEO–Chair trust and firm performance.
Dependent variables:	ROEt+1		NPMt+1		ROAt+1
	(1)	(2)	(3)	(4)	(5)	(6)
	OLS	OLS	OLS	OLS	OLS	OLS
TrustCEO,t	0.0072 **	0.0041 ***	0.0033 **	−0.0053	−0.0021	−0.0015
	(2.33)	(2.74)	(2.43)	(−1.29)	(−0.93)	(−0.76)
TrustChair,t				0.0136 ***	0.0066 ***	0.0050 ***
				(3.16)	(2.99)	(2.82)
Trustcompany,t	−0.0026	−0.0014	−0.0014	−0.0028	−0.0014	−0.0014
	(−0.87)	(−0.83)	(−0.98)	(−0.94)	(−0.82)	(−0.94)
Sizet	0.0532***	0.0286***	0.0238***	0.0558***	0.0298***	0.0248***
	(15.65)	(15.20)	(14.37)	(17.20)	(16.45)	(14.68)
FIN_levt	−0.0327	−0.0828***	−0.0436***	−0.0369	−0.0851***	−0.0457***
	(−1.23)	(−7.00)	(−3.97)	(−1.40)	(−7.28)	(−4.18)
Growth_ratet	0.0023***	0.0009**	0.0006	0.0022***	0.0009**	0.0005
	(2.85)	(2.17)	(1.62)	(2.89)	(2.15)	(1.58)
Firmaget	−0.0056	−0.0042	0.0009	−0.0052	−0.0038	0.0014
	(−1.16)	(−1.64)	(0.32)	(−1.09)	(−1.47)	(0.50)
BMt	−0.0229***	−0.0115***	−0.0119***	−0.0247***	−0.0121***	−0.0122***
	(−2.79)	(−3.60)	(−4.11)	(−3.08)	(−3.82)	(−4.21)
Volatilityt	0.0054	0.0040	0.0001	0.0034	0.0037	0.0001
	(0.52)	(0.74)	(0.02)	(0.34)	(0.68)	(0.02)
Boardsizet	0.0087	0.0016	0.0041	0.0136	0.0029	0.0045
	(0.58)	(0.19)	(0.58)	(0.98)	(0.36)	(0.63)
Dualityt	0.0185**	0.0107***	0.0086***	0.0206***	0.0109***	0.0086***
	(2.53)	(3.03)	(2.71)	(2.77)	(3.00)	(2.64)
Largestsharet	0.0551***	0.0120	0.0106	0.0564***	0.0117	0.0097
	(2.76)	(1.03)	(1.00)	(2.89)	(1.01)	(0.92)
Stateownt	−0.0182***	−0.0097**	−0.0098**	−0.0178**	−0.0095**	−0.0099**
	(−2.64)	(−2.34)	(−2.56)	(−2.53)	(−2.23)	(−2.47)
GDPgrowth,t	0.1738**	0.0649	0.0506	0.1796**	0.0664	0.0519
	(2.10)	(1.62)	(1.34)	(1.99)	(1.56)	(1.28)
Managerialownership,t	0.0531	0.0077	0.0143	0.0545	0.0085	0.0149
	(0.55)	(0.19)	(0.35)	(0.57)	(0.21)	(0.36)
Industry dummies	YES	YES	YES	YES	YES	YES
Year dummies	YES	YES	YES	YES	YES	YES
_Cons	−0.7006***	−0.3317***	−0.2816***	−0.7322***	−0.3446***	−0.2888***
	(−10.77)	(−10.08)	(−9.67)	(−12.03)	(−10.64)	(−9.73)
Obs.	2769	2769	2770	2676	2676	2677
Adjusted R2	0.3087	0.3933	0.3251	0.3235	0.4051	0.3338

Panel A presents the regression results of the impact of Chair–CEO trust on firm performance. The sample period is from 2000 to 2015 for the CEO–chair trust measures and control variables and from 2001 to 2016 for the firm performance measures. Panel B presents the regression results of the impact of CEO–Chair trust on firm performance over the sample period of 2000–2016 (columns 1–3) and the impacts of both Chair–CEO trust and CEO–Chair trust on firm performance over the sample period of 2000–2016 (columns 4–6). The t values reported in parentheses are based on standard errors clustered at both firm and year level. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively. All variables are defined in Appendix 3. OLS: ordinary least squares; IV-2SLS: instrumental variable–two-stage least squares.

Results reported in columns 2 to 7 of Panel A of Table 2 show that our baseline findings are unaffected after adding $T r u s t_{c o m p a n y}$ and using the 2SLS method. The coefficients of $T r u s t_{c h a i r}$ in the baseline regressions (see columns 2, 4, & 6) are statistically significant at the 5% or 1% levels. Not only are the effects statistically significant, but they are also economically meaningful. The increase of one standard deviation in Chair–CEO trust is associated with an increase of 1.61% (calculated as 0.0133 × 1.214) in ROE and an increase of 0.55% (i.e. 0.0045 × 1.214) in NPM, and 0.41% (calculated as 0.0034 × 1.214) in ROA.

The coefficients of control variables are mostly consistent with the literature. For example, firms with higher growth in operating incomes, of larger size, and located in a region with a higher GDP growth rate tend to perform better (Brush et al., 2000; Ferri, 2009; Shubita and Alsawalhah, 2012). Firms with greater stock return volatility, a higher BM, higher financial leverage, and state ownership tend to perform poorly (Chen and Zhang, 1998; Simerly and Li, 2000; Thomsen et al., 2006).

We next examine whether the other direction of trust, namely, the CEO’s trust in the management board chair, affects firm performance. Panel B of Table 2 presents our baseline regression results. In the test, the trust variable Trust_CEO is constructed similarly to $T r u s t_{c h a i r}$ , which is calculated using the provincial-level trustworthiness based on birthplaces of the CEO and the board chair. Repeating Model (1), we find a positive relationship between the CEO–Chair trust and firm performance. However, the CEO–Chair effect (b = 0.0072, column 1 of Panel B) is weaker than the Chair–CEO effect (b = 0.0133, column 2 of Panel A) on firm performance, evidenced by lower coefficient magnitudes (almost halved). The evidence is consistent with public opinion in China: the CEO is considered a subordinate of the board chair. Our evidence can be interpreted as the chair’s trust in the CEO dominating the studied relationship. Columns 3 to 6 of Panel B also report the regression results by including both trust variables ( $T r u s t_{C h a i r}$ and Trust_CEO). The results find that, when both directions of trust are considered, the Chair–CEO trust is still highly statistically significant, whereas the coefficient of the CEO–Chair trust becomes statistically insignificant in explaining firm performance. The evidence again suggests that Chair–CEO trust plays a more critical role than CEO–Chair trust in the studied relationship.

3.2. Advisory needs and advising quality

Following the literature’s suggestion that the board’s advice is more beneficial for firms with greater advisory needs (Kang et al., 2018; Schmidt, 2015), we use firm diversification levels to proxy for a firm’s advisory needs.⁷ Diversified firms engage in more complex businesses and are likely to have more significant advisory needs. Therefore, we predict that the positive relationship between Chair–CEO trust level and firm performance will be more pronounced in segmented firms than concentrated firms.

We separate segmented firms from concentrated firms using Diversified, which is equal to 1 if a firm operates in more than one CSRC (China Securities Regulatory Commission) two-digit code industry. We create an interaction term $D i v e r s i f i e d \times T r u s t_{c h a i r}$ and include the interaction term as well as Diversified in Regression Model (1). Results reported in Table 3 (columns 1–3, Panel A) show that Diversified is negatively correlated with firm performance. This negative correlation implies that firms operating in more than one segment perform worse than their counterparts (i.e. concentrated firms) in China (Singh and Gaur, 2009). Importantly, we find a consistent positive relationship of the interaction term ( $D i v e r s i f i e d \times T r u s t_{c h a i r}$ ) on firm performance, that is, significant at the 5% and 1% levels. This relationship implies that having a management-friendly board (i.e. the Chair–CEO trust) is more beneficial for diversified firms than for concentrated firms. The literature also suggests that highly financially leveraged firms have higher advisory needs than low-leveraged firms. We adopt the same method and find evidence suggesting (as reported in columns 4–6 of Panel A, Table 3) that having a management-friendly board is more beneficial for high-leverage firms (in the top leverage quartile) than for low-leverage firms (in the bottom leverage quartile).

Table 3.

The effect of advisory needs and advice quality.

Panel A: The effects of high- versus low-advisory needs.
	Segmented vs. concentrated firms			High-debt vs. low-debt firms
	(1)	(2)	(3)	(4)	(5)	(6)
	ROEt+1	NPMt+1	ROAt+1	ROEt+1	NPMt+1	ROAt+1
TrustChair,t	0.0023	0.0004	−0.0006	0.0092*	0.0028	0.0024
	(0.42)	(0.14)	(−0.27)	(1.90)	(1.55)	(1.40)
Diversifiedt × TrustChair,t	0.0172 ***	0.0062 **	0.0062 **
Diversifiedt × TrustChair,t	(2.88)	(2.46)	(2.66)
Diversifiedt	−0.0400***	−0.0188***	−0.0180***
	(−2.73)	(−2.83)	(−2.86)
Highdebtt × TrustChair,t				0.0163	0.0051 *	0.0027
Highdebtt × TrustChair,t				(1.62)	(1.85)	(1.13)
Highdebtt				−0.0140	−0.0065	−0.0060
				(−0.74)	(−0.92)	(−0.92)
Lowdebtt × TrustChair,t				0.0021	0.0021	0.0018
Lowdebtt × TrustChair,t				(0.38)	(0.55)	(0.50)
Lowdebtt				−0.0379	−0.0108	−0.0107
				(−1.38)	(−1.17)	(−1.26)
CONTROLS	YES	YES	YES	YES	YES	YES
Year dummies	YES	YES	YES	YES	YES	YES
Industry dummies	YES	YES	YES	YES	YES	YES
_cons	−0.9803***	−0.3495***	−0.2674***	−0.9593***	−0.3488***	−0.2657***
	(−7.72)	(−9.14)	(−7.39)	(−8.82)	(−8.80)	(−7.21)
Obs.	2418	2418	2419	2418	2418	2419
Adjusted R2	0.2422	0.4235	0.3521	0.2435	0.4196	0.3470

Panel B: The effect of advice quality.
	(1)	(2)	(3)	(4)	(5)	(6)
	ROEt+1		NPMt+1		ROAt+1
	Hqualadvice	Lqualadvice	Hqualadvice	Lqualadvice	Hqualadvice	Lqualadvice
TrustChair,t	0.0111 ***	0.0142	0.0057 ***	0.0033	0.0042 ***	0.0028
	(2.80)	(1.47)	(3.00)	(1.62)	(2.65)	(1.39)
CONTROLS	YES	YES	YES	YES	YES	YES
Year dummies	YES	YES	YES	YES	YES	YES
Industry dummies	YES	YES	YES	YES	YES	YES
_cons	−0.9298***	−1.0022***	−0.4173***	−0.3294***	−0.3519***	−0.2361***
	(−7.25)	(−4.96)	(−7.26)	(−7.54)	(−6.82)	(−5.50)
Obs.	1070	1348	1070	1348	1070	1349
Adjusted R2	0.3689	0.2270	0.4286	0.4428	0.3669	0.3703

Panel A presents the regression results of the impacts of business segmentation and debt level on the relationship between Chair–CEO trust and firm performance in China from 2000–2016. Panel B presents the regression results of the impact of advisory quality on the relationship between Chair–CEO trust and firm performance on the high-quality advising subsample (columns 1, 3, and 5) and the low-quality advising subsample (columns 2, 4, and 6). The t values reported in parentheses are based on standard errors clustered at both firm and year level. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively. All variables are defined in Appendix 3.

Field et al. (2013) suggest that busy boards (i.e. directors who sit on multiple corporate boards) have greater expertise derived from the experience of a variety of other companies’ board meetings and are better able to deliver high-quality advice. Field et al. (2013) argue that busy boards must efficiently allocate their attention to different companies and may thus focus on their advisory role at the expense of their monitoring role. Following Field et al. (2013), we define busy boards as being composed of directors who sit on more than three corporate boards. We split our sample into two groups: Hqual_advice and $L q u a l_{a d v i c e}$ . We define Hqual_advice firms as those with a higher than the sample mean ratio of busy directors (i.e. the number of busy boards divided by board size), and $L q u a l_{a d v i c e}$ contains the rest of the sample. We then repeat Regression Model (1) on the two advising quality subsamples. If our baseline result is correct, we would expect the relationship between $T r u s t_{c h a i r}$ and our three measurements of firm performance to be stronger in $H q u a l_{a d v i c e}$ firms. As predicted, results reported in Panel B of Table 3 show that the relationship between $T r u s t_{c h a i r}$ and firm performance is positive and highly significant (all at the 5% significance levels) in $H q u a l_{a d v i c e}$ firms (columns 1, 3, and 5). The effect attenuates in $L q u a l_{a d v i c e}$ firms, as evidenced by smaller coefficients of $T r u s t_{c h a i r}$ , which is statistically insignificant. This result is consistent with our prediction that the relationship between Chair–CEO trust level and firm performance is stronger in firms with boards that are more likely to offer high-quality advice.

3.3. Robustness

3.3.1. Addressing the concern of reverse causality: a placebo test

To verify whether our baseline result is affected by reverse causality, in that high-performing companies seek more trustworthy CEOs or board chairs, we design a placebo test that randomly assigns a Chair–CEO pair to a sample firm-year. Since the trust values are random, the null of no impact on firm performance should be observed. We record 1000 estimates of point coefficients on simulated samples using Model (1), with ROE as the dependent variable. Our simulation results (illustrated in Figure 2) show that the coefficients of $T r u s t_{c h a i r}$ range from −0.0069 to 0.0099, with a mean of 0.00002. Based on these simulation results, there is a zero in 1000 chances of randomly observing the coefficient (0.0139) reported in column 2 of Table 2 when the null of no effect of Chair–CEO trust on firm performance holds. Relating the simulation results to our baseline test, the t values in Table 2 are conservative. We also repeat the placebo test using ROA and NPM as the dependent variables, respectively, to proxy for firm performance, and the same conclusions hold.

Figure 2.

Placebo tests of the effect of Chair–CEO trust on firm performance measured by return on equity (ROE).

3.3.2. Addressing the concern of omitted variable bias

3.3.2.1. Control for educational/political background, overseas working experience, and CEO–Chair duality

The baseline model does not consider the observed characteristics of the board chair or the CEO. To ensure that our results are not driven by their inherent traits, experience, or expertise, we add a set of variables to account for their age, gender, overseas study/work experience, tenure, highest educational qualifications, and political background in the regression model. This step follows Chen (2015) and Giannetti et al. (2015): $C E O_{a g e}$ , $C E O_{g e n d e r}$ , CEO_oversea, CEO_tenure, $C E O_{e d u}$ , $C E O_{p o l i t i c a l}$ , Chair_age, Chair_gender, Chair_oversea, $C h a i r_{e d u}$ , $C h a i r_{p o l i t i c a l}$ . These variables are defined in Appendix 3. Similar to the disclosure of birthplace information, the disclosure of educational background and overseas experience is voluntary; our sample size drops to 1477 after we exclude missing values for these additional variables. Panel A in Table 4 presents the results with this set of controls (denoted as “CONTROLS”) in addition to the same set of control variables presented in Table 2. We find that the positive relationship of the Chair–CEO trust on firm performance remains unchanged after controlling for the CEO’s and the chair’s characteristics.

Table 4.

Address the concern for omitted variable bias.

Panel A: Characteristics of the CEO and board chair.
	(1)	(2)	(3)
	ROEt+1	NPMt+1	ROAt+1
TrustChair,t	0.0073 **	0.0033 **	0.0024
	(1.98)	(1.99)	(1.65)
CEOpolitical	0.0140*	0.0041	0.0027
	(1.90)	(1.04)	(0.75)
Chairpolitical	−0.0112	−0.0052	−0.0039
	(−1.46)	(−1.30)	(−1.06)
CEOoversea	0.0016	−0.0088	−0.0094
	(0.08)	(−1.30)	(−1.21)
Chairoversea	−0.0212	−0.0029	−0.0014
	(−0.89)	(−0.40)	(−0.16)
CEOgender	−0.0161	−0.0071	−0.0068
	(−1.19)	(−0.99)	(−1.05)
Chairgender	−0.0063	−0.0040	−0.0018
	(−0.39)	(−0.51)	(−0.23)
CEOedu	−0.0044	0.0003	0.0001
	(−0.96)	(0.12)	(0.04)
Chairedu	0.0072	0.0016	0.0011
	(1.41)	(0.59)	(0.48)
CEOtenure	−0.0001	−0.0001	−0.0001
	(−0.78)	(−1.16)	(−0.94)
Chairtenure	0.0002*	0.0001	0.0001
	(1.93)	(1.58)	(1.37)
CEOage	0.0029	0.0140	0.0125
	(0.14)	(1.16)	(1.14)
Chairage	−0.0207	−0.0222	−0.0140
	(−0.81)	(−1.43)	(−1.00)
CONTROLS	YES	YES	YES
YEAR	YES	YES	YES
INDUSTRY	YES	YES	YES
_cons	−0.5381***	−0.2411***	−0.2339***
	(−3.45)	(−3.08)	(−3.30)
Obs.	1477	1477	1477
Adjusted R2	0.3870	0.4568	0.3806

Panel B: Control for the impacts of education resources and economic developments in the metropolitan cities of Beijing and Shanghai.
	(1)	(2)	(3)
	ROEt+1	NPMt+1	ROAt+1
TrustChair,t	0.0133 **	0.0040 ***	0.0030 **
	(2.06)	(2.73)	(2.28)
Metropoliset	−0.0010	0.0062	0.0039
	(−0.05)	(0.71)	(0.50)
CONTROLS	YES	YES	YES
Year dummies	YES	YES	YES
Industry dummies	YES	YES	YES
_cons	−0.9607***	−0.3543***	−0.2958***
	(−8.20)	(−10.76)	(−9.74)
Obs.	2812	2812	2813
Adjusted R2	0.2263	0.4001	0.3208

Panel C: CEO and the chair’s alumni relationship.
	(1)	(2)	(3)
	ROEt+1	NPMt+1	ROAt+1
TrustChair,t	0.0144 **	0.0046 ***	0.0036 **
	(2.22)	(2.74)	(2.34)
Alumni	−0.0419***	−0.0188***	−0.0225***
	(−3.52)	(−3.81)	(−4.68)
Missingalumni	−0.0175	−0.0005	0.0007
	(−1.64)	(−0.11)	(0.16)
CONTROLS	YES	YES	YES
Year dummies	YES	YES	YES
Industry dummies	YES	YES	YES
__cons	−0.9667***	−0.3557***	−0.2737***
	(−7.95)	(−8.72)	(−7.13)
Obs.	2418	2418	2419
Adjusted R2	0.2380	0.4170	0.3445
Panel D: Exclude cases that both a company’s CEO and its chair are from the same province.
	(1)	(2)	(3)
	ROEt+1	NPMt+1	ROAt+1
TrustChair,t	0.0272 **	0.0088 ***	0.0066 ***
	(2.11)	(3.44)	(2.59)
CONTROLS	YES	YES	YES
Year dummies	YES	YES	YES
Industry dummies	YES	YES	YES
_cons	−0.8751***	−0.2109***	−0.1507**
	(−3.05)	(−3.31)	(−2.47)
Obs.	749	749	749
Adjusted R2	0.2637	0.4401	0.3615

Panel E: Exclude local chairs or local CEOs.
Dependent variables:	ROEt+1		NPMt+1		ROAt+1
	(1)	(2)	(3)	(4)	(5)	(6)
	Difchair	DifCEO	Difchair	DifCEO	Difchair	DifCEO
TrustChair,t	0.0227 *	0.0191 **	0.0062 **	0.0058 **	0.0045 *	0.0043 *
	(1.93)	(2.07)	(2.37)	(2.31)	(1.84)	(1.78)
CONTROLS	YES	YES	YES	YES	YES	YES
Year dummies	YES	YES	YES	YES	YES	YES
Industry dummies	YES	YES	YES	YES	YES	YES
_cons	−0.8038***	−0.7977***	−0.1927***	−0.2075***	−0.1405**	−0.1504***
	(−3.34)	(−3.36)	(−3.12)	(−3.33)	(−2.49)	(−2.60)
Obs.	848	839	848	839	848	839
Adjusted R2	0.2522	0.2453	0.4179	0.4240	0.3408	0.3516

Panel F: Including both the post-turnover (the successors) and pre-turnover CEO–Chair trust in explaining firm performance.
	(1)	(2)	(3)
	ROEt+1	NPMt+1	ROAt+1
CEOturnover	−0.0971**	−0.0241***	−0.0198***
	(−2.01)	(−3.49)	(−2.84)
Trustpost	0.0325 *	0.0075 ***	0.0061 ***
	(1.85)	(3.74)	(2.69)
Trustpre	0.0066	0.0015	0.0009
	(1.24)	(1.04)	(0.76)
CONTROLS	YES	YES	YES
Year dummies	YESYES	YES	YES
Industry dummies	YESYES	YES	YES
_cons	−0.7797***	−0.3090***	−0.2623***
	(−5.35)	(−7.38)	(−6.94)
Obs.	2035	2035	2036
Adjusted R2	0.2253	0.4184	0.3422

Panel G: The impact of CEO–Chair trust on firm performance in the event of CEO turnovers.
	(1)	(2)	(3)
	ROEt+1	NPMt+1	ROAt+1
TrustChair,t	0.0080	0.0017	0.0009
	(1.27)	(1.05)	(0.61)
CEOturnover	−0.0835**	−0.0197***	−0.0167**
	(−2.00)	(−2.95)	(−2.47)
Trustchairt × CEOturnover	0.0255 *	0.0059 ***	0.0052 **
Trustchairt × CEOturnover	(1.85)	(2.87)	(2.32)
CONTROLS	YES	YES	YES
Year dummies	YES	YES	YES
Industry dummies	YES	YES	YES
_cons	−0.8329***	−0.3255***	−0.2528***
	(−5.44)	(−7.86)	(−6.39)
Obs.	1959	1959	1960
Adjusted R2	0.2363	0.4336	0.3544

Panel A presents the regression results of the impact of Chair–CEO trust on firm performance in China from 2000 to 2016, after including the chair’s and CEO’s backgrounds. This panel presents the regression results of the impact of Chair–CEO trust on firm performance after controlling for economic growth and educational resources for Beijing-born and Shanghai-born CEOs and board chairs from 2000 to 2016. Panel C presents the regression results of the impact of Chair–CEO trust on firm performance after controlling for the CEO and chair’s alumni relationship in China from 2000 to 2016. Panel D presents the regression results of the impact of Chair–CEO trust on firm performance after excluding cases that both a company’s CEO and its chair are from the same province from 2000 to 2016. In Panel E, columns 1, 3, and 5 present the regression results of the impact of Chair–CEO trust on firm performance in firms whose chair is not from the local province where the firm is headquartered (denoted as Difchair, consisting of firms with non-local chairs only). Columns 2, 4, and 6 present the regression results of the impact of Chair–CEO trust on firm performance in firms whose CEO is not from the local province where the firm is headquartered (denoted as DifCEO, consisting of firms with non-local CEOs only). Panel F presents the regression results of the impact of Chair–CEO trust by using both the post-turnover (the successors) and pre-turnover CEO–chair trust to explain firm performance in China from 2000 to 2016. Panel G presents the impact of CEO–chair trust on firm performance in the event of CEO turnovers in China from 2000–2016. The t values reported in parentheses are based on standard errors clustered at the levels of firm and year. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively. All variables are defined in Appendix 3.

3.3.2.2. Control for the impacts of educational resources and economic developments

A concern may arise that economic growth and educational resources can enforce trustworthiness in the regions and that these regions are also likely to foster more businessman. In other words, if economic developments and educational resources can fully explain the trust index, after considering this relationship, our trust–performance relationship will disappear. We test this possibility in this section.⁸

The provincial trust data in Appendix 2.1 show that Shanghai and Beijing are the two cities with the highest trustworthy levels in the country. Appendix 2.2 finds that 6.12% of the chairs and 7.07% of the chairs in our sample are from Beijing and Shanghai, respectively. The two largest metropolitan cities in northern and eastern China have the best educational resources and the highest GDP growth rate in the country. Even before the founding of the People’s Republic of China, Beijing and Shanghai had been the most developed cities in the nation. To implement the new test, an indicator variable of Metropolis is created and included in the baseline regression model. We assign the value of 1 to Metropolis when the chairman and the CEO are from Beijing or Shanghai, and 0 otherwise. The test results presented in Panel B of Table 4 find that the coefficient of $T r u s t_{c h a i r}$ remains significant while the coefficient of Metropolis is insignificant. This indicates that local economic growth and educational resources cannot explain our baseline regression finding.

3.3.2.3. Control for the relationships of alumni between the chair and the CEO

In this test, we control for the effect of alumni relationship. We capture the alumni relationship between the CEO and the chair with a dummy variable of Alumni. The variable takes the value of 1 when the CEO and the chair graduated from the same university, and 0 otherwise. Again, the positive relationship between Chair–CEO trust and firm performance persists after controlling for the alumni relationship (Panel C of Table 4).

Appointment decisions of the chair and the CEO may confound the trust effect of driving firm performance. For example, the chair may nominate a CEO from the same region when they are personally related. To verify this competing interpretation, we drop observations where the CEO and the chair in a company were born in the same province. We find the positive effect of the Chair–CEO trust on firm performance is even stronger, as evidenced by higher coefficient values for $T r u s t_{c h a i r}$ (Panel D of Table 4) than those reported in Table 2.

3.3.2.4. Control for other social capitals of the board chair and the CEO

Another potential endogeneity concern is that the positive relationship between Trust_chair and firm performance may be driven by some unobservable social connection factors. As companies prefer to hire local CEOs or chairs, a substantial portion of our sample are firms with local chairs and local CEOs (with birthplaces in the same province as the firm’s headquarters location). Firms with local CEOs or chairs may benefit from this social capital and produce better firm performance. For example, Adler and Kwon (2002) propose that entrepreneurs may access resources through their networks. Such social capital can allow entrepreneurs to identify investment opportunities (Bhagavatula et al., 2010) and build legitimacy for their firms (Elfring and Hulsink, 2003). Political connections allow firms to secure favorable regulatory conditions, which can also result in improved firm performance (Agrawal and Knoeber, 2001; Johnson and Mitton, 2003). In this case, the positive relationship between CEO trust and firm performance may be driven by the social capital of local chairs and local CEOs.

To address this concern, we first show that firms with local CEOs and local chairs do not always have a higher level of Chair–CEO trust. Using the trust index table in Appendix 1, we find that people sometimes more greatly trust those from a more developed region than theirs. For example, the trust level of people from Hebei, Liaoning, and Heilongjiang relating to people from Beijing is 27.7, 22.6, and 18.4, respectively, higher than the level at which people living in these areas trust their fellow residents (17, 20.8, and 11.8, respectively). This evidence suggests that firms with local CEOs and local chairs do not always exhibit a higher Chair–CEO trust level.

To exclude the effect of local CEOs or local chairs in a formal test, we assemble two subsamples: (1) a subsample named $D i f_{C E O}$ that consists of observations with CEOs who are not from the province where the company is headquartered, and (2) a subsample named $D i f_{c h a i r}$ that consists of observations with chairs who are not from the province where the company is headquartered. Using the two subsamples helps mitigate the concern that our result is driven by observing situations where the CEO and the chair come from the same region. We do not construct a subsample excluding observations where either the chair or the CEO is local because the sample size is too small, which may hinder the reliability of the regression results. Panel E of Table 4 shows that the significant and positive relationship of the Chair–CEO trust on firm performance is consistent across these subsamples.

3.3.2.5. Exploiting the event of CEO turnovers

When a CEO is newly appointed to a firm, they typically require a high level of advice and information exchange from the board to rapidly acclimatize and adapt them to their new position. The event of CEO turnovers also provides an opportunity to test whether the trust of the foregoing CEO still plays a role in explaining firm performance after the CEO turnover. One argument is that Chair–CEO trust may change the corporate culture that affects firm performance to a certain extent even after his or her departure. Econometrically, this test also helps alleviate a concern of endogeneity caused by omitted variables. That is, if our result is driven by some omitted variables that simultaneously determine the selection decision of CEO and firm performance, the pre-turnover Chair–CEO trust can continue to significantly impact firm performance. To test, we construct two regression models as follows

\begin{array}{l} F P_{i, t + 1} = β_{0} + β_{1} C E O_{t u r n o v e r} + β_{2} T r u s t_{p o s t} + β_{3} T r u s t_{p r e} \\ + γ C o n t r o l s_{i t} + I n d u s t r y + Y e a r + ε_{i, t + 1} \end{array}

Model (2)

\begin{array}{l} F P_{i, t + 1} = β_{0} + β_{1} T r u s t_{C h a i r, i t} + β_{2} C E O_{t u r n o v e r, i t} + β_{3} C E O_{t u r n o v e r, i t} \times T r u s t_{c h a i r, i t} \\ + γ C o n t r o l s_{i t} + I n d u s t r y + Y e a r + ε_{i, t + 1} \end{array}

Model (3)

In the above models, $C E O_{t u r n o v e r}$ is an indicator of CEO turnover. If the company changes its CEO in the current year t, the value is 1. To test whether firm performance is driven by chair–successor CEO trust (Trust_{chair, post}) or chair–predecessor CEO trust ( $T r u s t_{c h a i r, p r e}$ ) in CEO turnovers, we include both variables in Model (2). Note, the variable of $T r u s t_{c h a i r, p o s t}$ is identical to $T r u s t_{c h a i r}$ that we use in our baseline regression. In Model (3), we add the indicator of CEO turnover, $T r u s t_{c h a i r}$ , and their interaction term, $C E O_{t u r n o v e r} \times T r u s t_{c h a i r}$ , to the baseline regression.

Panel F of Table 4 reports the results of Model (2). Consistent with the literature (Kato and Long, 2006), we find a negative relationship between $C E O_{t u r n o v e r}$ and our three performance measures. We show that only current Chair–CEO trust matters (the coefficients of $T r u s t_{p o s t}$ are positive and statistically significant in all the three columns); the effect of previous trust disappears after a new relationship is formed, as evidenced by the statistical insignificance of the coefficient of $T r u s t_{p r e}$ . That is, we find firm that performance is not driven by chair–predecessor CEO trust after the CEO turnover. Results reported in Panel G of Table 4 complement the results in Panel F that Chair–CEO trust is beneficial to firms: the positive and significant coefficients of $C E O_{t u r n o v e r} \times T r u s t_{c h a i r}$ indicate that Chair–CEO trust can moderate the negative impact of CEO turnover on firm performance.⁹

4. What channels through which does Chair–CEO trust drive firm performance?

The literature suggests that management-friendly boards can improve firm performance by improving advisory consequences. Below we consider the implications of two strategic business decisions where firms are most likely to benefit from a robust advisory process, namely, corporate innovation and MA.

4.1. Trust and innovation performance

A company’s innovation activities usually involve a multi-stage, long-term process to develop and eventually commercialize new and untested ideas (Holmstrom, 1989). Such innovation typically requires significant board advisory inputs (Kang et al., 2018). As such, it is crucial for management to discuss the innovation process with the board so that it can provide timely feedback about strategies for both the current and next stage of development (Manso, 2011). In addition, the innovation process requires expertise and firm-specific knowledge. Therefore, communication between the board and management is important for the board to provide quality advice. Since innovation is significant to achieving competitive advantage (Yuan and Wen, 2018), which requires board advisory input, we propose that a friendly board can improve firm performance by motivating firm innovation. Based on the context we researched, we propose that high Chair–CEO trust can accelerate the information exchange that enables the board to provide quality advice that is crucial for value-adding innovation activities. These activities of a firm can be assessed in terms of its input (R&D investment) and output (the quality/value of patents received, for example, Chen et al., 2013; Yuan and Wen, 2018). As data on patent citations, an important metric that measures patent quality, are unavailable in China (Yuan and Wen, 2018), we thus assess only the effect of Chair–CEO trust on R&D in determining firm performance. Following the literature, we construct a measure of R&D inputs using annual R&D increase over total assets. Model (4) below is used to test this mechanism

\begin{matrix} R & D_{i, t + 1} = β_{0} + β_{1} H i g h_{c h a i r, i t} + β_{2} H i g h_{c h a i r, i t} \times R & D_{i t} + β_{3} L o w_{c h a i r, i t} \\ + β_{4} L o w_{c h a i r, i t} \times R & D_{i t} \\ + β_{5} R & D_{i t} + γ C o n t r o l s_{i t} + I n d u s t r y + Y e a r + ε_{i, t + 1} \end{matrix}

Model (4)

We employ the same set of control variables of firm characteristics used in Model (1), including firm- or stock-level controls (i.e. $T r u s t_{c o m p a n y}$ , Size, Fin_lev, BM, Firm_age, Growth_rate, Volatility), corporate governance and ownership controls (Boardsize, Stateown, LargestInstitution, $M a n a g e r i a l_{o w n e r s h i p}$ , and Duality), and regional economic growth rate ( $G D P_{g r o w t h}$ ). We also include the variables used by Yuan and Wen (2018) as additional control variables: $C E O_{a g e}, R O A,$ $C a s h_{r a t i o}, A s s e t_{t u r n o v e r},$ $G r o w t h_r a t e$ , $I n s t i t u t i o n a l$ , $B u s y_{d i r e c t o r}$ , and Managerial_ownership. Appendix 3 defines these variables. The regression models also include industry-fixed effects and year-fixed effects.

We create two indicators to capture high versus low Chair–CEO trust level: $H I G H_{c h a i r}$ and $L O W_{c h a i r}$ . They are assigned a value of 1 when the value of $T r u s t_{c h a i r}$ is in the top quartile/the bottom quartile, and 0 otherwise. We interact the two indicators with the R&D variable. We add the two indicator variables as well as the interaction terms in Regression Model (4).

If CEO–chair trust indeed influences R&D investment decisions that, in turn, impact on firm performance, we predict that the significant and positive relationship is stronger in $H I G H_{c h a i r} \times R & D$ and weaker in $L O W_{c h a i r} \times R & D$ . The results show that the coefficient of $H I G H_{c h a i r} \times R & D$ is persistently significant as predicted (see columns 1, 3, and 5, Panel A of Table 5), at the 10% or 5% levels. The effect is even more significant for technology firms as it relates to coefficient magnitude and significance level (columns 2, 4, and 6, Panel A of Table 5). In contrast, the coefficient of $L O W_{c h a i r} \times R & D$ is consistently insignificant according to both the overall sample and the subsample results. This evidence indicates that corporate R&D investment decision-making is a channel through which Chair–CEO trust drives firm performance.

Table 5.

Mechanisms whereby Chair–CEO trust drives firm performance through strategic decision-making in Research and Developments (R&D) and Merger and Acquisitions (MA).

Panel A: The joint effects of Chair–CEO trust and R&D investments on firm performance.
Dependent variables:	ROEt+1		NPMt+1		ROAt+1
	(1)	(2)	(3)	(4)	(5)	(6)
	Full sample	High tech	Full sample	High tech	Full sample	High tech
HIGHChair	0.0184	0.0028	0.0054	0.0057	0.0057	0.0036
	(1.29)	(0.14)	(1.33)	(0.51)	(1.56)	(0.35)
HIGHChair × R&D	0.0132 **	0.0148 *	0.0101 **	0.0111 **	0.0083 **	0.0095 **
	(2.14)	(1.95)	(2.19)	(2.33)	(2.16)	(2.45)
LowChair	−0.0096	−0.0213	−0.0047	−0.0087	−0.0025	−0.0036
	(−1.26)	(−1.24)	(−1.28)	(−0.95)	(−0.71)	(−0.41)
LowChair × R&D	−0.0040	0.0021	0.0004	0.0058	−0.0005	0.0040
	(−0.75)	(0.23)	(0.16)	(1.10)	(−0.19)	(0.83)
R&D	−0.0002	−0.0048	−0.0007	−0.0016	−0.0003	−0.0015
	(−0.06)	(−0.67)	(−0.36)	(−0.37)	(−0.20)	(−0.40)
CONTROLS	YES	YES	YES	YES	YES	YES
Year dummies	YES	YES	YES	YES	YES	YES
Industry dummies	YES	YES	YES	YES	YES	YES
_cons	−0.9152***	−0.7308***	−0.3313***	−0.3217***	−0.2521***	−0.2418***
	(−8.22)	(−4.89)	(−9.11)	(−4.24)	(−7.08)	(−3.16)
Obs.	2526	464	2526	464	2527	464
Adjusted R2	0.2273	0.4080	0.4098	0.4832	0.3365	0.3915

Panel B: Univariate test results on the merger and acquisition sample.
	Obs.	M	t value	Duality
(1) CAR1 [0, 1]	151	0.003	2.13**	0
(2) CAR1 [0, 1]	60	−0.002	−0.71	1
(3) CAR3 [0, 3]	151	0.007	2.18**	0
(4) CAR3 [0, 3]	60	0.003	0.85	1
(5) CAR5 [0, 5]	151	0.006	1.33	0
(6) CAR5 [0, 5]	60	0.007	1.32	1

Panel C: The relationship between Chair–CEO trust and cumulative abnormal returns on merger and acquisition (MA) announcements in the event windows of [0, 1], [0, 3], and [0, 5].
Dependent variables:	CAR1		CAR3		CAR5
	(1)	(2)	(3)	(4)	(5)	(6)
	MA sample	Duality = 0	MA sample	Duality = 0	MA sample	Duality = 0
TrustChair,t	0.0052 ***	0.0033	0.0091 ***	0.0144 ***	0.0141 **	0.0211 ***
	(3.36)	(1.08)	(2.82)	(2.64)	(1.99)	(2.96)
CONTROLS	YES	YES	YES	YES	YES	YES
Year dummies	YES	YES	YES	YES	YES	YES
Industry dummies	YES	YES	YES	YES	YES	YES
_cons	0.0905	0.1007	−0.0253	−0.0432	−0.1182**	−0.1682***
	(1.41)	(1.61)	(−0.38)	(−0.38)	(−2.20)	(−3.00)
Obs.	193	135	193	135	193	135
Adjusted R2	0.4228	0.5268	0.4221	0.6064	0.3952	0.5825

Panel A presents the regression results on the joint effects of Chair–CEO trust and R&D investments on firm performance in China from 2000 to 2016. Columns 1, 3, and 5 present the regression results on the full sample. Columns 2, 4, and 6 present the regression results on technology firms only (denoted as High tech, consisting of firms in the industry of information technology). Panel B presents the results of market reactions to sample stocks with merger and acquisition announcements during the sample period of 2000–2016. Rows 1, 3, and 5 present the CARs in the three event windows of [0, 1], [0, 3], and [0, 5] around the MA announcements in our sample of MA firms where a CEO does not take the dual position of chair (Duality = 0). Rows 2, 4, and 6 present the CARs on the merger and acquisition announcements in the three event windows of [0, 1], [0, 3] and [0, 5] for the MA firms where a CEO also takes the position of a chair (Duality = 1). Panel C presents the results of the impact of Chair–CEO trust on firm acquisition performance in China from 2000 to 2016. Columns 1, 3, and 5 present the CARs for the merger and acquisition announcements in the three event windows of [0, 1], [0, 3] and [0, 5] for all the MA sample firms. Columns 2, 4, and 6 present the CARs for the MA firms where a CEO does not take the position of management board chair (Duality = 0). ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively. The t values reported in parentheses are based on standard errors clustered at both firm and year level. CARs: cumulative abnormal returns.

4.2. Trust and market reactions to the announcement of MA

Masulis and Mobbs (2011) suggest that acquisition involves an important board decision that can directly affect firm performance because it requires a “thorough understanding of the expected synergies and costs of the acquisition and the risks associated with the transaction.” They suggest that better-informed boards can assess MA more accurately and therefore improve its performance. Because acquisition decisions are part of a board’s typical advisory function (Baldenius et al., 2014; Kim et al., 2014), and if Chair–CEO trust can enhance firm value by improving this advice, MA events may exemplify the effect of Chair–CEO trust. Using this argument, we predict a positive relationship between Chair–CEO trust level and MA performance.

We measure MA performance using cumulative abnormal returns (CARs) estimated on market reactions to the MA announcements in 1-day, 3-day, and 5-day event windows (0, 1), (0, 3), and (0, 5) (denoted as $C A R_{1}$ , $C A R_{3}$ , and $C A R_{5})$ . We use the market model to estimate CARs, with the value-weighted stock index as the market return. We next test the relationship between Chair–CEO trust and CARs using Model (5) below

C A R_{i t} = a + β_{1} T R U S T_{c h a i r, i t} + γ C o n t r o l s_{i t} + I n d u s t r y + Y e a r + ε_{i t}

Model (5)

where the dependent variable is the CARs on the [0, 1], [0, 3], and [0, 5] announcement windows. We conduct this model using the MA sample (rather than the full sample in the main analysis) and the MA sample excluding duality cases (where duality = 0). To avoid the sample size being too small to perform a regression, in this section, we code observations with missing chair/CEO birthplace information using the following procedures. First, if a firm is missing the birthplace information of its chair, the Chair–CEO trust is calculated as the natural logarithm of the average cross-provincial-level trustworthiness based on the birthplace of the CEO. For example, according to Appendix 1, the average cross-provincial-level trustworthiness of people from Beijing is 19.6. If in our sample, a company has a CEO from Beijing, but its chair’s birthplace information is missing, $T r u s t_{c h a i r}$ is then assigned a value of 2.98 (i.e. the natural logarithm of 19.6).

Second, if a firm has a missing value of the CEO’s birthplace, the Chair–CEO trust is calculated as the natural logarithm of the average cross-provincial-level trustworthiness based on birthplace of the chair. For example, according to Appendix 1, people from Beijing express average cross-provincial-level trustworthiness of 6.68. If, in our sample, a company has a chair from Beijing and is missing the CEO’s birthplace information, the value of Trust_chair is 1.9 (i.e. the natural logarithm of 6.68).

Third, if a firm has missing values regarding both the chair’s and CEO’s birthplace, $T r u s t_{c h a i r}$ takes the value of 1.79, the mean value of $T r u s t_{c h a i r}$ . In addition, we include a disclosure indicator variable, $M i s s i n g$ , which takes the value of 1 if the birthplace information of a company’s board or its CEO is undisclosed, and 0 otherwise. The control variables are the same as per Model (1).

Panel B of Table 5 reports the univariate analysis for our two MA samples: (1) firms where the CEO also takes the chair position in the management board (i.e. duality = 1), and (2) firms where the CEO does not take the position of management board chair (i.e. duality = 0). Univariate analysis is conducted to test whether MA performance is related to a board’s advisory role. A CEO, who is also the board chair, may not value or follow the advice of other directors because his or her power is superior. Therefore, if the MA performance is related more to the board’s advisory role, we expect non-duality firms to have better acquisition performance. Results in Panel B of Table 5 support this expectation. We find that the short-term MA return is positive for non-duality firms in the 1-day, 3-day, and 5-day event windows (both the 1-day and 3-day announcements are statistically significant), but insignificant for duality firms.

Panel C of Table 5 reports the results of Model (3). Columns 1, 3, and 5 report the regression results on the full MA sample and columns 2, 4, and 6 contain only non-duality firms. As predicted, the coefficients of Chair–CEO trust and MA performance are all positive and statistically significant in most cases. Combined with our main finding that Chair–CEO trust can increase firm value, these results suggest that MA decision-making is another channel through which Chair–CEO trust drives firm performance.

5. Conclusion

In this article, we provide empirical evidence that the trust between the board chair and the CEO (i.e. Chair–CEO trust) is essential to a sound advisory board process. Specifically, trust promotes valuable information sharing and collaboration, thereby driving firm performance. The positive effect of trust remains unchanged after controlling for the impacts of regional economic growth, industry, firm, and personal characteristics. Additional test results suggest that the relationship is causal. The trust–performance relationship is more pronounced in firms with more significant advisory needs and boards that can deliver high-quality advice. The study also explores the results of corporate R&D and MA, two strategic business decisions that may benefit from sound advisory procedures. These results support our view that high degrees of chairman–CEO trust can increase company value.

Footnotes

Appendix

Appendix 3.

Variable definitions.

Firm performance variables (Data source: CSMAR)		Variables appear in the tables
ROA	The sum of total profits and financial expenses divided by the average balance of total assets.	Tables 1 –5
ROE	Net profit divided by the average balance of shareholder’s equity.	Tables 1 –5
NPM	Net profits divided by the average balance of total assets.	Tables 1 –5
Trust variables (Source: Zhang and Ke (2002))
TrustChair	The natural logarithm of the trust score as a measure of the trust of the board chair in the CEO of a company, with the trust score sourced from Appendix 1. For example, as shown in Appendix 1, people from Shanghai evaluate the trustworthiness level of people from Beijing as 7.4 (row 9, column 2). In our context, if a company has a chair from Shanghai and a CEO from Beijing, the TrustChair is 2 (ln(7.4) = 2).	Tables 1 –5
TrustCEO	The natural logarithm of the trust score as a measure of the trust of the CEO in the board chair of a company, with the trust score sourced from Appendix 1. For example, as shown in Appendix 1, people from Shanghai evaluate the trustworthiness level of people from Beijing as 7.4 (row 9, column 2); people from Beijing evaluate the trustworthiness level of people from Shanghai as 23.9 (row 2, column 10). If in our sample, a company has a CEO from Shanghai and a chair from Beijing, the value of TrustCEO is 2 (ln(7.4) = 2).	Table 1
Control variables: firm characteristics (Source: Zhang and Ke (2002) and CSMAR)
TrustCompany	Natural logarithm of the trust score according to the headquarters location province of a company, calculated as per Zhang and Ke (2002).	Tables 1; Tables 2 –5 in CONTROLS
FIN_lev	The ratio of total liabilities to total assets.	Tables 1; Tables 2 –5 in CONTROLS
BM	The book value of equity divided by the market value of equity.	Tables 1; Tables 2 –5 in CONTROLS
Size	The natural logarithm of current year market capitalization.	Tables 1; Tables 2 –5 in CONTROLS
Firmage	The natural logarithm of firms’ existing years since the IPO year.	Tables 1; Tables 2 –5 in CONTROLS
Growth_rate	The difference between total operating income in the current year and total operating income in the same period of the previous year divided by total operating income in the same period of previous year.	Tables 1; Tables 2 –5 in CONTROLS
Volatility	The standard deviation of (logarithm) daily returns in the previous 250 trading days.	Tables 1; Tables 2 –5 in CONTROLS
IPO: initial public offering.
Control variables: corporate governance and ownership structure (Source: CSMAR)
Duality	A dummy variable that equals 1 if the CEO also takes the position of the board chair in the same company, and 0 otherwise.	Tables 1; Tables 2 –5 in CONTROLS
Boardsize	The total number of board members.	Tables 1; Tables 2 –5 in CONTROLS
Other control variables (Source: Wind)
Stateown	Equal to 1 if the company is a state-owned, and 0 otherwise.	Tables 1; Tables 2 –5 in CONTROLS
Largestsharet	Number of shares owned by the largest shareholder divided by total shares outstanding in the current year.	Tables 1; Tables 2 –5 in CONTROLS
GDP_Growth	The natural logarithm of GDP per capita for each province at the firm-year level.	Tables 1; Tables 2 –5 in CONTROLS
GDP: gross domestic product.
CEO and chair background (Source: CSMAR)
CEOoverseas	A dummy variable that is equal to 1 if the CEO has undertaken overseas study or working experience.	Table 4
CEOage	The natural logarithm of a CEO’s age.	Table 4
CEOtenure	The number of years that the CEO has been in his or her position as a CEO.	Table 4
CEOpolitical	A dummy variable that is equal to 1 if the CEO has ever been a government official, including a member of the National Provincial People’s Congress, or the provincial People’s Congress, or the Municipal People’s Congress, or Chinese People’s Political Consultative Conference; and 0 otherwise.	Table 4
CEOedu	CEOs’ educational background, where 1 = receive no more than senior school education, 2 = receive junior college education, 3 = with bachelor’s degree, 4 = with master’s degree, 5 = with PhD or higher degree.	Table 4
CEOgender	A dummy variable that is equal to 1 if the CEO’s gender is male, equal to 0 if the CEO’s gender is female.	Table 4
Chairage	The natural logarithm of a chair’s age.	Table 4
Chairoverseas	A dummy variable that is equal to 1 if the chair has overseas study or working experience.	Table 4
Chairtenure	The number of years that the chair has been in his or her position as a chair.	Table 4
Chairpolitical	A dummy variable that is equal to 1 if the CEO has ever been a government official, including a member of the National Provincial People’s Congress, or the provincial People’s Congress, or the Municipal People’s Congress, or Chinese People’s Political Consultative Conference; and 0 otherwise.	Table 4
Chairedu	Chairs’ educational background, where 1 = receive no more than senior school education, 2 = receive junior college education, 3 = with bachelor’s degree, 4 = with master’s degree, 5 = with PhD or higher degree.	Table 4
Chairgender	A dummy variable that is equal to 0 if the chair’s gender is male, equal to 1 if the chair’s gender is female.	Table 4
CEO turnovers (Source: CSMAR)
CEOturnover	An indicator of CEO turnover event, which takes the value of one if a firm changes its CEO in the current year.	Table 4
Trustpre	Chair–CEO trust level before the CEO turnover.	Table 4
Trustpost	Chair–CEO trust level after the CEO turnover.	Table 4
Chair-CEO ties (Source: CSMAR)
Alumni	A dummy variable that takes the value of 1 if the CEO and the chair graduated from the same university, and 0 otherwise.	Table 4
Missingalumni	A dummy variable that takes the value of 1 if the firm does not disclose the chair’s or the CEO’s educational background, and 0 otherwise.	Table 4
Research & Investment, M&A variables (Source: CSMAR)
HIGHChair	A dummy variable that is equal to 1 if the value of TRUSTChair is in the top quantile.	Table 5
LowChair	A dummy variable that is equal to 1 if the value of TRUSTChair is in the bottom quantile.	Table 5
R&D	The net increase in the current year R&D expenditure scaled by total assets.	Table 5
Cashholdingt	Cash or cash equivalent divided by total assets.	Table 5
%Busydirector	The percentage of directors who participate in more than three companies divided by the total directors.	Table 5
CAR1	The cumulated abnormal return over the 1-day window [0, 1] around the MA announcement dates.	Table 5
CAR3	The cumulated abnormal return over the 3-day window [0, 3] around the MA announcement dates.	Table 5
CAR5	The cumulated abnormal return over the 5-day window [0, 5] around the MA announcement dates.	Table 5

Final transcript accepted 19 November 2020 by Philip Gray (AE Finance).

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Yushu Zhu

Notes

References

Adams

Ferreira

(2007) A theory of friendly boards. Journal of Finance 62: 217–250.

Adler

Kwon

(2002) Social Capital: Prospects for a new concept. The Academy of Management Review 27: 17–40.

Agrawal

Knoeber

(2001) Do some outside directors play a political role? The Journal of Law and Economics 44: 179–198.

Anderson

Reeb

(2004) Board composition: Balancing family influence in S&P 500 firms. Administrative Science Quarterly 49: 209–237.

Bai

Liu

, et al. (2004) Corporate governance and market valuation in China. Journal of Comparative Economics 32: 599–616.

Baldenius

Melumad

Meng

(2014) Board composition and CEO power. Journal of Financial Economics 112: 53–68.

Baliga

Moyer

Rao

(1996) CEO duality and firm performance: What’s the fuss? Strategic Management Journal 17: 41–53.

Beck

Frost

Jones

(2018) CSR disclosure and financial performance revisited: A cross-country analysis. Australian Journal of Management 43: 517–537.

Bhagavatula

Elfring

van Tilburg

, et al. (2010) How social and human capital influence opportunity recognition and resource mobilization in India’s handloom industry. Journal of Business Venturing 25: 245–260.

10.

Bonin

Huang

(2002) Foreign entry into Chinese banking: Does WTO membership threaten domestic banks? The World Economy 25: 1077–1093.

11.

Booth

Deli

(1996) Factors affecting the number of outside directorships held by CEOs. Journal of Financial Economics 40: 81–104.

12.

Brush

Bromiley

Hendrickx

(2000) The free cash flow hypothesis for sales growth and firm performance. Strategic Management Journal 21: 455–472.

13.

Champion

(1999) Finance: The joy of leverage. Harvard Business Review 77: 19–22.

14.

Chen

Guan

Zhang

, et al. (2017) Political connection of financial intermediaries: Evidence from China’s IPO market. Journal of Banking & Finance 76: 15–31.

15.

Chen

Liu

Wang

(2016) Social trust and bank loan financing: Evidence from China. Abacus 52: 374–403.

16.

Chen

Zhang

(1998) Risk and return of value stocks. Journal of Business 71: 501–535.

17.

Chen

Liang

, et al. (2013) R&D spillover effects and firm performance following R&D increases. Journal of Financial and Quantitative Analysis 48: 1607–1634.

18.

Chen

(2015) Institutions, board structure, and corporate performance: Evidence from Chinese firms. Journal of Corporate Finance 32: 217–237.

19.

Claessens

Djankov

Lang

(2000) The separation of ownership and control in East Asian Corporations. Journal of Financial Economics 58: 81–112.

20.

Coleman

(1990) Foundations of Social Theory. Cambridge, MA: Belknap Press of Harvard.

21.

Coles

Daniel

Naveen

(2008) Boards: Does one size fit all? Journal of Financial Economics 87: 329–356.

22.

Dahya

Karbhari

Xiao

(2002) The supervisory board in Chinese listed companies: Problems, causes, consequences and remedies. Asia Pacific Business Review 9: 118–137.

23.

Daily

Dalton

(1994) Bankruptcy and corporate governance: The impact of board composition and structure. Academy of Management Journal 37: 1603–1617.

24.

Deng

Jiang

, et al. (2019) Social capital and access to informal finance —Evidence from Chinese private firms. Accounting & Finance 59: 2767–2815.

25.

Dong

Han

, et al. (2016) Social trust and corporate misconduct: Evidence from China. Journal of Business Ethics 151: 539–562.

26.

Elfring

Hulsink

(2003) Networks in entrepreneurship: The case of high-technology firms. Small Business Economics 21: 409–422.

27.

Fama

French

(1998) Value versus growth: The international evidence. Journal of Finance 53: 1975–1999.

28.

Farag

Mallin

(2017) Monitoring corporate boards: Evidence from China. The European Journal of Finance 25: 524–549.

29.

Ferri

(2009) Are new tigers supplanting old mammoths in China’s banking system? Evidence from a sample of city commercial banks. Journal of Banking and Finance 33: 131–140.

30.

Field

Lowry

Mkrtchyan

(2013) Are busy boards detrimental? Journal of Financial Economics 109: 63–82.

31.

Garrett

Hoitash

Prawitt

(2014) Trust and financial reporting quality. Journal of Accounting Research 52: 1087–1125.

32.

Giannetti

Liao

(2015) The brain gain of corporate boards: Evidence from China. Journal of Finance 70: 1629–1682.

33.

Guiso

Sapienza

Zingales

(2004) The role of social capital in financial development. American Economic Review 94: 526–556.

34.

Habib

Bhuiyan

MBU

Hasan

(2018) Firm life cycle and advisory directors. Australian Journal of Management 43: 575–592.

35.

Hadlock

James

(2002) Do banks provide financial slack? Journal of Finance 57: 1383–1419.

36.

Harris

Raviv

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

37.

Hillman

Shropshire

Certo

, et al. (2011) What I like about you: A multilevel study of shareholder discontent with director monitoring. Organization Science 22: 675–687.

38.

Holmstrom

(1989) Agency costs and innovation. Journal of Economic Behavior and Organization 12: 305–327.

39.

Hovey

Naughton

(2003) The relationship between valuation and ownership of listed firms in China. Corporate Governance 11: 112–122.

40.

Jha

Chen

(2015) Audit fees and social capital. The Accounting Review 90: 611–639.

41.

Jia

Bradbury

(2020) Risk management committees and firm performance. Australian Journal of Management 8: 120–140.

42.

Jiang

Kim

(2015) Corporate governance in China: A modern perspective. Journal of Corporate Finance 32: 190–216.

43.

Jiang

Cai

Nofsinger

, et al. (2020) Can reputation concern restrain bad news hoarding in family firms? Journal of Banking & Finance 114: 105808.

44.

Jin

Wang

, et al. (2016) Social trust and foreign ownership: Evidence from qualified foreign institutional investors in China. Journal of Financial Stability 23: 1–14.

45.

Johnson

Mitton

(2003) Cronyism and capital controls: Evidence from Malaysia. Journal of Financial Economics 67: 351–382.

46.

Jones

Makri

Gomez-Mejia

(2008) Affiliate directors and perceived risk bearing in publicly traded, family–controlled firms: The case of diversification. Entrepreneurship Theory and Practice 32: 1007–1026.

47.

Kang

Liu

Low

, et al. (2018) Friendly boards and innovation. Journal of Empirical Finance 45: 1–25.

48.

Kato

Long

(2006) Executive turnover and firm performance in China. American Economic Review 96: 363–367.

49.

Khan

Mather

Balachandran

(2014) Managerial share ownership and operating performance: Do independent and executive directors have different incentives? Australian Journal of Management 39: 47–71.

50.

Kim

Mauldin

Patro

(2014) Outside directors and board advising and monitoring performance. Journal of Accounting and Economics 57: 110–131.

51.

Klein

(1998) Firm performance and board committee structure. The Journal of Law and Economics 41: 275–304.

52.

LaPorta

Lopez-de-Silanes

Shleifer

, et al. (1997) Trust in large organizations. American Economic Review Papers and Proceedings 87: 333–338.

53.

Lee

(2009) Does size matter in firm performance? Evidence from US public firms. International Journal of the Economics of Business 16: 189–203.

54.

Naughton

(2007) Going public with good governance: Evidence from China. Corporate Governance: An International Review 15: 1190–1202.

55.

Wang

(2017) Trust and stock price crash risk: Evidence from China. Journal of Banking & Finance 76: 74–91.

56.

Lin

Lee

Hung

(2006) R&D intensity and commercialization orientation effects on financial performance. Journal of Business Research 59: 679–685.

57.

Liu

Luo

Tian

(2016) Managerial professional connections versus political connections: Evidence from firms’ access to informal financing resources. Journal of Corporate Finance 41: 179–200.

58.

Manso

(2011) Motivating innovation. Journal of Finance 66: 1823–1860.

59.

Masulis

Mobbs

(2011) Are all inside directors the same? Evidence from the external directorship market. Journal of Finance 66: 823–872.

60.

Matolcsy

Wright

(2011) CEO compensation structure and firm performance. Accounting & Finance 51: 745–763.

61.

Park

Lee

(2020) The sensitivity of corporate social performance to corporate financial performance: A “time-based” agency theory perspective. Australian Journal of Management. Epub ahead of print 31 May. DOI: 10.1177/0312896220917192.

62.

Putnam

(1993) The prosperous community: Social capital and public life. The American Prospect 13: 5–42.

63.

Qiu

Chan

(2019) Does social trust restrain firm financing violations? Evidence from China. Accounting & Finance. Epub ahead of print 10 December. DOI: 10.1111/acfi.12583. Available at: https://onlinelibrary-wiley-com.ezproxy.library.uq.edu.au/doi/epdf/10.1111/acfi.12583 (accessed 3 July 2020).

64.

Quigley

Hambrick

(2011) When the former CEO stays on as board chair: Effects on successor discretion, strategic change, and performance. Strategic Management Journal 33: 834–859.

65.

Rechner

Dalton

(1991) CEO duality and organizational performance: A longitudinal analysis. Strategic Management Journal 12: 155–160.

66.

Roden

Lewellen

(1995) Corporate capital structure decisions: Evidence from leveraged buyouts. Financial Management 24: 76–87.

67.

Rose

Shepard

(1997) Firm diversification and CEO compensation: Managerial ability or executive entrenchment? The RAND Journal of Economics 28: 1–45.

68.

Schmidt

(2015) Costs and benefits of friendly boards during mergers and acquisitions. Journal of Financial Economics 117: 424–447.

69.

Shleifer

Vishny

(1997) A survey of corporate governance. Journal of Finance 52: 737–783.

70.

Shubita

Alsawalhah

(2012) The relationship between capital structure and profitability. International Journal of Business and Social Science 3: 104–112.

71.

Simerly

(2000) Environmental dynamism, capital structure and performance: A theoretical integration and an empirical test. Strategic Management Journal 21: 31–49.

72.

Singh

Gaur

(2009) Business group affiliation, firm governance, and firm performance: Evidence from China and India. Corporate Governance: An International Review 17: 411–425.

73.

Stevenson

Radin

(2009) Social capital and social influence on the board of directors. Journal of Management Studies 46: 16–44.

74.

Thomsen

Pedersen

Kvist

(2006) Blockholder ownership: Effects on firm value in market- and control-based governance systems. Journal of Corporate Finance 12: 246–269.

75.

Wang

(2003) Governance role of different types of state-shareholders: Evidence from China’s listed companies. Unpublished PhD Dissertation, Hong Kong University of Science and Technology, Kowloon, Hong Kong.

76.

Firth

Rui

(2014) Trust and the provision of trade credit. Journal of Banking & Finance 39: 146–159.

77.

Xia

Cao

Chan

(2017) Social trust environment and firm tax avoidance: Evidence from China. The North American Journal of Economics and Finance 42: 374–392.

78.

Xiao

Dahya

Lin

(2004) A grounded theory exposition of the role of the Supervisory Board in China. British Journal of Management 15: 39–55.

79.

Yermack

(1996) Higher market valuation of companies with a small board of directors. Journal of Financial Economics 40: 185–211.

80.

Yuan

Wen

(2018) Managerial foreign experience and corporate innovation. Journal of Corporate Finance 48: 752–770.

81.

Zhang

(2002) Trust in China: A cross-regional analysis. Economic Research Journal 10: 59–70.