Introduction

Because CEOs may act opportunistically (Fama and Jensen 1983; Jensen and Meckling 1976), directors are called upon to monitor the CEO on behalf of shareholders and other stakeholders (Fama 1980). They also provide resources, such as skills and connections to other organizations that may enhance organizational performance (Carpenter and Westphal 2001; Pfeffer and Salancik 1978). Although one might expect that all board members will monitor and provide resources assiduously in order to fulfill their obligations, Westphal and colleagues (e.g., Westphal and Stern 2006; Westphal and Zajac 1997) suggest that the interpersonal relationship between the CEO and directors has important implications in terms of how well the directors discharge their roles. Because trust affects all interpersonal relationships (Dirks and Ferrin 2001), we focus on how a director’s perceptions of the CEO’s trustworthiness are expected to affect his or her resource provision and monitoring behaviors.

In this paper, we therefore examine directors’ monitoring and resource provision from a relational perspective. Specifically, we address the following research questions: (1) how do a director’s perceptions of the CEO’s trustworthiness affect the director’s governance behaviors, and (2) how do the board practices (i.e., board performance evaluation practices) moderate the relations between the director’s trustworthiness perceptions and his or her governance behaviors. We draw on Mayer et al. (1995) multidimensional view of trustworthiness (i.e., ability, integrity, and benevolence) to argue that a director’s perceptions of the CEO will affect his or her resource provision and monitoring. We also argue that board performance evaluation practices will interact with the director’s perceptions of the CEO’s trustworthiness to affect the director’s behaviors. These board practices will likely affect the director’s incentives to fulfill his or her board duties. Our focus in this study is at the individual director level, rather than the board level, so we examine the interpersonal relationships between each director and the CEO. Our empirical findings drawn from data from three different institutional contexts (i.e., Canada, Singapore, Spain) show that a director’s perceptions of the CEO’s trustworthiness are related to his or her monitoring and resource provision.

One important contribution of this paper is that we introduce the concept of trust between a director and the CEO, and present a model that links the trustworthiness perceptions to individual directors’ governance behaviors. We empirically show that a director’s perceptions of the different dimensions of the CEO’s trustworthiness have different effects on the director’s behaviors. We demonstrate that positive trustworthiness perceptions of integrity and benevolence are associated with greater resource provision consistent with the argument that higher trust is a positive factor that brings about greater cooperation (Dirks and Ferrin 2001; Westphal 1999) and we show that positive trustworthiness perceptions of integrity lead to reduced monitoring (Langfred 2004; McEvily et al. 2003). Regarding ability, while the results are not consistent with our hypotheses, they show evidence of a non-linear relationship with monitoring and with resource provision, which indicates that directors only provide resources when the perceptions of the CEO’s ability are low. At high levels of ability perceptions, directors seem to believe that the CEO is able to perform his or her role appropriately without their cooperation.

Regarding monitoring, directors monitor the CEO for most levels of ability perceptions, and they only reduce monitoring for extremely high levels of perceived ability.

The second contribution of our paper is that we examine the effects of the performance evaluation practices of the board. Although such practices are becoming increasingly common, we still know little about the impact of such practices on directors’ governance behaviors. Therefore, we explain how the frequency of board evaluations (e.g., of board members and the board’s effectiveness) may affect the director’s monitoring and resource provision.

Directors’ Governance Behaviors

Directors of the board typically play two different yet equally important roles: monitoring CEOs and providing resources to them (Adams and Ferreira 2007; Hillman and Dalziel 2003; Hillman et al. 2008). Monitoring is emphasized in agency theory literature, which assumes that managers may act opportunistically (Jensen and Meckling 1976). The monitoring role therefore includes assessing the CEO’s performance, monitoring strategy implementation, designing CEO compensation schemes, and CEO succession planning (Boyd 1995), and it can result in the dismissal of the CEO for poor performance (Daily and Dalton 1995).

Resource dependence research has focused on directors’ expertise, knowledge, and skills as well as their ties to external organizations and their effects on organizational performance. Hillman and Dalziel (2003) argue that this human and relational capital is the antecedent of the directors’ resource provision activities. The focus of resource dependence research is on the board members’ abilities rather than their motivations to provide valuable resources to the firm (Daily et al. 2003). However, researchers have begun to consider the directors’ resource provision and incentives concurrently, by integrating agency and resource dependence theories (Hillman and Dalziel 2003). We build on these advances by considering new antecedents of each of these behaviors.

The monitoring and resource provision roles are qualitatively different. While both functions demand high levels of expertise, how the director’s expertise is utilized differs; monitoring requires the application of expertise in information processing and assessment, while resource provision entails offering resources (e.g., external ties and valuable knowledge) in a concrete fashion (e.g., Pfeffer and Salancik 1978). In this paper, following recent studies on the role of the board that consider monitoring and resource provision separately (Adams and Ferreira 2007; Hillman and Dalziel 2003; Westphal 1999, Acero and Alcalde 2012), we also regard these as two separate activities.

Trust in the Corporate Governance Context

Trust has not been widely studied in the agency theory literature on corporate governance because agency theory assumes a lack of trust between parties in its examination of the risk of managerial opportunism and conflicts of interests. In contrast, stewardship theory (Davis et al. 1997; Donaldson 1990) and stakeholder theory (Jones 1995) assume greater trust in management and goal alignment between parties. By comparing two models of board governance, one predicated on trust-based collaboration, and the other predicated on distrust-based control (or risk of opportunism), Sundaramurthy and Lewis (2003) suggest that greater trust-based collaboration between the CEO and the board can lead to complacency and entrenchment, while greater control (which signals distrust) leads to even further control and a division between the CEO and the board. This research suggests that the CEO-board relationship can be characterized by either the presence or absence of trust, and trust is treated as a unidimensional construct. Our model presents detailed mechanisms of how varying levels of a director’s perceptions of each dimension of the CEO’s trustworthiness may affect the director’s resource provision and monitoring.

Although the board functions as a group, each individual director develops his or her own subjective perceptions of the CEO’s trustworthiness and acts accordingly. Although we acknowledge the effects of the board norms on individual directors, we focus on the perceptions and behaviors of individual directors. Trustworthiness perceptions could potentially be measured at the board level, but it is likely that the level of such perceptions varies significantly among directors for various reasons, making aggregation difficult (Kozlowski and Klein 2000). For example, some directors may attribute poor firm performance to the CEO’s past decisions or his/her implementation of strategy, while others may attribute it to external factors that are beyond the CEO’s control (e.g., the economy). These differences in attributions may lead to very different perceptions of the CEO’s ability. Also, some directors may have social ties with the CEO and have interactions outside the boardroom, which can affect how they assess the CEO’s benevolence and integrity. We therefore contend that individual directors often have varying levels of trustworthiness perceptions of the CEO, and that it is more instructive to examine the effects of these individual perceptions. We now turn to interpersonal trust.

Trust in Interpersonal Relationships

Interpersonal interactions in organizations are generally governed by an unspoken social exchange between individuals, and the fundamental driver of this exchange relationship is trust (Blau 1964; Holmes 1981; Homans 1958; Chiaburu and Lim 2008). Trust usually develops over time, as reciprocal obligations are met and as both parties have an opportunity to observe the other’s behaviors (Blau 1964). Rousseau et al. (1998, p. 395) define trust as “a psychological state comprising the intention to accept vulnerability based upon positive expectations of the intentions or behavior of another”. Mayer et al. (1995, p. 712) similarly define trust as the “willingness to be vulnerable to another party when that party cannot be controlled or monitored”. These definitions imply that trusting another entails risk and vulnerability. Trust is an integral component of effective social exchanges in organizations, and has been linked to risk-taking, task performance, and organizational citizenship behaviors (Chou et al. 2008; Colquitt et al. 2007; Dirks 1999). We believe that trust must also affect the behaviors of corporate directors.

Trust, however, is not a monolithic construct. Mayer et al. (1995) present three separate dimensions of trustworthiness: the other party’s ability, benevolence, and integrity. “Ability” (sometimes termed ability-based trust, Brockner and Siegel 1996; cognition-based trust, McAllister 1995; competence trust, Lui 2009; or task reliability, Sitkin and Roth 1993) stems from a perception that the trustee has the professional knowledge, skills and abilities to fulfill his or her required tasks. “Integrity” (sometimes labeled intent-based trust, Brockner and Siegel 1996) is defined as “perception that the trustee adheres to a set of principles that the trustor finds acceptable” (Mayer and Gavin 2005, p. 874). “Benevolence” is defined by Mayer et al. (1995) as the belief that the trustee cares about the well-being of the potential trustor. Previous research has tested this 3D model empirically (Aubert and Kelsey 2003; Jarvenpaa and Leidner 1999; Yakovleva et al. 2010) and shown that the different types of trust have different consequences (e.g., Kim et al. 2004). In this paper, we apply the 3D model of trustworthiness to the director–CEO relationship. In our view, a director’s perceptions of the CEO’s ability, integrity, and benevolence, or their perceptions of the CEO’s trustworthiness, will affect the director’s likelihood of engaging in resource provision or monitoring.

The CEO’s ability can be defined as the extent to which his or her skills and competence enable him or her to perform the requisite managerial duties. Perceptions about the CEO’s abilities are task-specific; our interest here is the CEO’s capacity to perform his or her managerial tasks, keeping in mind that several different skill sets and knowledge may be required (e.g., financial, marketing, leadership). Perceptions of the CEO’s integrity are based on the extent to which the director believes that the CEO follows acceptable principles and values, in terms of how they are reflected in the CEO’s honest, consistent, and open behavior (Serva and Fuller 2004). Perceptions of the CEO’s benevolence are determined by the extent to which the director believes that the CEO wants to “do well by” the director. Benevolence implies a specific personal relationship (Mayer et al. 1995), and this suggests that the CEO may be benevolent to one director but not to another. Our interest lies in the CEO’s benevolence toward the focal director.

Researchers have suggested that ability is the most objective of the three trustworthiness dimensions and is the most important in a particular task-context, such as the management of the firm (e.g., Serva and Fuller 2004). Perceptions of the CEO’s ability (e.g., making strategic decisions) are formed by observing and by gathering other information. Because integrity and benevolence perceptions are assessed more subjectively, the relational context plays a role in shaping those perceptions (Yakovleva et al. 2010). In our view, perceptions of integrity and benevolence provide key information for assessing the other party’s propensity to reciprocate and to avoid acting opportunistically or unethically (Cruz et al. 2010).

In typical exchange relationships, in which the norm of reciprocity is critical because the “giver” of resources cannot require the recipient to reciprocate (Blau 1964; Gouldner 1960), the giver has to decide whether to risk the recipient’s possible opportunistic behavior or non-reciprocity. In the context of the director–CEO relationship, we suggest that integrity and benevolence perceptions affect how the director assesses the risk of the CEO acting opportunistically and of his or her provision of resources not being reciprocated by the CEO. In other words, benevolence and integrity perceptions are more likely to affect directors’ social exchange motivations. For example, in response to the director’s provision of valuable market knowledge or a linkage with an external party that may greatly benefit the firm, the CEO may reciprocate by not using the given resources for personal benefit, by supporting the director’s appointment at another board where the CEO sits as a director,Footnote 1 or by formally acknowledging the director’s contributions, which will enhance the director’s reputation. Social exchange theory allows us to link the concept of trust, which implies vulnerability and risk, with directors’ governance behaviors.

We expect that the extent of monitoring activities will also be affected by directors’ perceptions of the potential risks to the company (and to the director) due to any perceived deficiencies in the CEO’s trustworthiness, especially ability and integrity. As benevolence perceptions are relevant in specific individual relationships (i.e., one can be benevolent to one individual but not to another individual), the potential risks and benefits are also related to specific individuals. Shen (2003) argues that the intensity of board monitoring will vary according to the risk of a CEO’s opportunistic behavior. These risks may include the possibility of the CEO engaging in opportunistic or unethical behaviors such as hiding unfavorable information or negative results, or making incompetent or even unapproved strategic decisions. Similarly, the extent of resource provision activities will likely be affected by the potential risks to the company due to perceived deficiencies in the CEO’s ability and integrity. Directors may be concerned that the CEO will act opportunistically and use the resources provided by the director for unintended purposes. These risks are potentially problematic to the firm and to the directors, because the directors are responsible for overseeing managerial decisions and protecting the interests of the firm; any negative consequences due to inappropriate CEO behaviors will also jeopardize the directors’ reputation. Therefore, each director is expected to use some discretion in how he or she behaves, and it is the director’s assessment of the CEO’s trustworthiness and the potential risks that affect how much monitoring and resource provision the director chooses to do.

As noted by Mayer and Davis (1999), there may be instances where an evaluation of one dimension (i.e., ability) may differ from that of another (i.e., benevolence). We have therefore considered the effects of a director’s perceptions of the CEO’s ability, benevolence, and integrity on the director’s governance behaviors separately. It is equally critical to emphasize that each of these dimensions are evaluated along a continuum. Although it would be rare for a CEO to possess a very low level of ability, integrity, or benevolence because this would result in him or her being removed, there will necessarily be some variation in the extent to which a CEO possess each of these qualities, relative to each other, and relative to his or her peers. Furthermore, each of these qualities is perceived by directors who are assessing the CEO in relation to their own expectations of how CEOs should act.

It is important to note that while we examine the effects of a director’s trustworthiness perceptions of a CEO, such perceptions can be affected by how the CEO perceives the director’s trustworthiness as well. As suggested by Ferrin et al. (2008), the levels of trust between individuals often evolve over time through repeated interactions. Hence, as in many other interpersonal relationships, the director–CEO relationship may change over time. For example, a high level of monitoring by a director may lead to greater social distance between the director and the CEO, which in turn may lead to lower levels of the director’s perceptions of the CEO’s benevolence and hence even more monitoring. While these dynamic relationships are possible in the director–CEO relationship, our study focuses on the effects of trustworthiness perceptions on the director’s behaviors because our interest in this study is to extend the trust research, which often examines the relationships between trust and cooperative and monitoring behaviors in other work settings, to the relationship within the boardroom.Footnote 2

The Effect of Ability Perceptions on Monitoring and Resource Provision

Because the CEO’s strategic decisions can have important implications for organizational performance, it is reasonable to expect that board members will assess the CEO’s ability carefully. Among the three dimensions of trustworthiness, the ability perceptions are more likely to invoke concerns about firm performance rather than social exchange risks because, as we discussed, the ability is the most objective measure among the three dimensions (Serva and Fuller 2004) and will likely have a direct impact on performance. We argue that a director’s perceptions of the CEO’s ability will have a negative relation with his or her monitoring, with greater monitoring activity occurring when the CEO has especially low levels of perceived ability. With positive perceptions of the CEO’s ability, the director will be more likely to risk being vulnerable to the CEO’s actions and decisions, because he or she has confidence that the CEO is capable of performing the necessary managerial duties. This logic is consistent with the trust research that treats trust as a substitute of monitoring. Therefore, the director may allow the CEO greater discretion and autonomy, and de-emphasize monitoring and control (Langfred 2004; McEvily et al. 2003).

In contrast, if a director has negative perceptions of the CEO’s ability to perform the required duties, he or she will likely not feel comfortable taking a risk by granting great autonomy to the CEO; the director would be more likely to, once again, intensify his or her monitoring of the CEO (Bromiley and Cummings 1995). One possible outcome of intense monitoring is CEO dismissal; keeping an incompetent CEO would have potentially detrimental effects on organizational performance and also increase the risks to the individual directors (e.g., litigation). In cases where the lack of ability is not so severe as to merit termination (e.g., small deficiencies in some areas are overshadowed by significant strengths in more important areas), or if termination is not possible (e.g., family owned firms), the director will provide resources in order to reduce the risks caused by the CEO’s difficulties in managing the organization. In this case, the director would protect his or her own interests (e.g., prevent damage to his or her reputation from being associated with a poorly performing organization). It is also possible that the director would have ethical qualms about not supporting the CEO with resources, given the obligation to help the organization to succeed.

The effect of directors’ perceptions of the CEO’s ability on the director’s resource provision activities is expected to be similar. We argue that a director’s perceptions of the CEO’s ability will predict his or her resource provision activities, with higher levels of resources being provided to CEOs with low levels of perceived ability. The CEO’s use of the provided resources will potentially improve firm performance, thereby helping to enhance firm performance as well as the director’s reputation. The director would therefore be motivated to provide more resources when his or her perceptions of the CEO’s ability are positive. We therefore propose:

Hypothesis 1a

A director’s perceptions of the CEO’s ability have a negative relation to his or her monitoring activities; perceptions of low ability are related to more monitoring.

Hypothesis 1b

A director’s perceptions of the CEO’s ability have a negative relation to his or her resource provision activities; perceptions of low ability are related to more resource provision.

The Effect of Integrity Perceptions on Monitoring and Resource Provision

Although the CEO’s ability is an important determinant of directors’ governance behaviors, it is equally important to consider the CEO’s integrity. Negative perceptions of the CEO’s integrity may raise questions as to whether the CEO intends to act appropriately and honestly (Rousseau et al. 1998), especially since people tend to weigh negative information more heavily than positive information, when assessing others’ moral character (Snyder and Stukas 1999). The perceived lack of integrity of the CEO can also have negative performance implications to the firm, as the CEO’s unethical behavior may lead to some negative consequences. A director’s negative perceptions of the CEO’s integrity may therefore incline the director to place more emphasis on monitoring. Perceptions that the CEO’s integrity is adequate will provide reasonable reassurance that the CEO will act in a manner consistent with the director’s own principles and values. Hence, the director may be willing to risk being more vulnerable to the actions of the CEO, and engage in less monitoring, keeping in mind that monitoring is time consuming and at times difficult. Further, as noted by Gulati and Westphal (1999), monitoring behaviors have the potential to create social distance between a CEO and a director. Hence, from a social exchange perspective, the director is less motivated to intensify his or her monitoring when she or he perceives that the CEO has a high level of integrity.

The director’s perceptions of the CEO’s integrity may also affect his or her resource provision activities. As discussed earlier, relational context comes into play when one assesses another’s integrity. The director’s positive perceptions of the CEO’s integrity may lead to the director’s expectation of a reciprocal exchange relationship with the CEO. When resources are provided, the director may expect that the CEO with high integrity will respond and reciprocate in an honest, open, and consistent manner (Serva and Fuller 2004). For example, the CEO may respond by being forthcoming with negative or sensitive information, or by sharing more information with the director (Dirks and Ferrin 2001; Nahapiet and Goshal 1998). This kind of interaction creates a positive reciprocity in their social exchange relationship (Blau 1964; Molm 1994). When a director has positive perceptions of the CEO’s integrity, then, the director’s resource provision role is likely to be enhanced. On the other hand, negative perceptions may lead the director to question whether his or her resources would be used appropriately and honestly by the CEO. We therefore propose the following:

Hypothesis 2a

A director’s perceptions of the CEO’s integrity have a negative relation to his or her monitoring activities; negative perceptions of integrity are related to more monitoring.

Hypothesis 2b

A director’s perceptions of the CEO’s integrity have a positive relation to his or her resource provision activities; positive perceptions of integrity are related to more resource provision.

The Effect of Benevolence Perceptions on Monitoring and Resource Provision

As in any personal relationship, including those between a principal (i.e., a director who represents shareholders) and an agent (i.e., the CEO), perceptions of the other party’s benevolence, or the belief that the agent cares about the well-being of the principal (Mayer et al. 1995), should affect how one perceives the risk of opportunism (Cruz et al. 2010). When one perceives that the other party is benevolent toward oneself, one can expect that the other party will not act opportunistically or harm one’s interests; benevolence is the opposite of opportunism (Cruz et al. 2010; Schoorman et al. 2007; Yakovleva et al. 2010). Perceptions that the CEO has an adequate level of benevolence, then, may lead the director to lower his or her monitoring activities, because high levels of safeguards against opportunistic behaviors may not be necessary (McEvily et al. 2003). Greater monitoring can also create social distance and hence, the director who perceives the CEO to be benevolent will not likely wish to risk the social relationship with the CEO through intense monitoring.

A director’s perception of the CEO’s benevolence, or the belief that the CEO cares about the director’s well-being (Mayer et al. 1995), will also affect his or her resource provision. Positive perceptions of benevolence signal that the CEO is willing to “do well by” the director. From a social exchange perspective, positive benevolence perceptions suggest that the CEO would not take advantage of the director and would respond positively to the director’s resource provision. A director’s positive perceptions of the CEO’s benevolence, then, will lead the director to provide a benevolent CEO with resources, in the interests of promoting a positive relationship with an individual who may be disposed to reciprocate such treatment. However, in cases where the director’s perceptions of the CEO’s benevolence are negative, there are few incentives for the director to provide resources to the CEO. There would be high levels of uncertainty about whether this resource provision would be reciprocated by the CEO; as a result, the director would expose him/herself to a high risk of non-reciprocity. We therefore propose that:

Hypothesis 3a

A director’s perceptions of the CEO’s benevolence have a negative relation to his or her monitoring activities; positive perceptions of benevolence are related to less monitoring.

Hypothesis 3b

A director’s perceptions of the CEO’s benevolence have a positive relation to his or her resource provision activities; positive perceptions of benevolence are related to more resource provision.

The Effects of Board Performance Evaluation Practices

We have thus far focused on individual directors’ subjective assessments of the CEO, and how these affect the directors’ governance behaviors. However, board members’ performance is under increasing scrutiny and an increasing number of boards are conducting peer assessments and using them for the performance evaluations of individual directors. Because the board functions as a group (Golden and Zajac 2001) and individual board member’s behaviors are constantly assessed by other members, it is important to consider the effects of such board practices on individual directors’ behaviors. Although each board has its own expectations about how much effort each director should expend to perform his or her board duties (Lorsch and Macliver 1989; Zona and Zattoni 2007), the presence of such board practices will likely increase directors’ monitoring and resource provision, because they will be more aware of negative consequences that may accrue to a failure to perform their duties adequately. We therefore make the following hypothesis:

Hypothesis 4

The presence of board performance evaluation practices is positively related to a director’s monitoring and resource provision.

In terms of the moderating effects of board performance evaluation practices, it is likely that they will positively interact with a director’s perceptions of the trustworthiness of the CEO. Because we predict that the board performance evaluation practices have positive effects on the director’s monitoring and resource provision, we expect that the presence of such practices will further enhance the likelihood that directors will monitor and provide resources. We therefore present the following hypothesis:

Hypothesis 5

The presence of board performance evaluation practices positively moderates the relationship between a director’s perceptions of the CEO’s ability, benevolence, and integrity and his or her monitoring and resource provision.

Method

Sample and Data Collection

Our study focuses on a sample of 160 outside directors coming from three different countries: Spain, Singapore and Canada. Data were collected through an anonymous survey questionnaire sent to outside directors of major organizations in the three countries. We chose these geographical contexts to enhance generalizability of our findings. These countries have different cultural and institutional characteristics. However, the official responsibilities of the board are very similar. Our target population includes directors of both profit and non-profit organizations. Although the relative emphasis of the board activities is arguably slightly different between these two types of organizations, board members are still expected to monitor management and provide resources. Therefore, our study focuses on behaviors of the director as a profession.

The survey was undertaken from October 2009 to March 2010. Formal agreements were obtained when possible with directors associations in the different countries in order to facilitate the implementation of the study. It was especially effective in Spain where a tight relationship was established with the Institute of Directors of Spain (IC-A). Members of these associations and institutes include directors of most major firms, family firms and non-profit organizations in the country. A cover letter was included in all cases explaining the relevance of the study and survey feedback was promised to be delivered both to the associations and to all those directors who provided us with their email addresses. The first reminder was sent after 3 weeks after and the second reminder was sent 6 weeks after the fist reminder. After removing responses with missing values, the final number of responses for the three countries was thus 160 observations: 60 responses from Singapore, 70 responses from Canada, and 30 usable responses from Spain.

Of these 160 respondents who indicated where their company was registered, 44 % were Canadian, 38 % were from Singapore, and 18 % were from Spain. Of the 72 respondents who indicated their gender, most were male (71 %). In our sample, the average age of the 92 respondents who indicated it was 56. This is comparable with the average age of boards of directors in Singapore (50; Quah 2006), Spain (59; Spencer Stuart Board Index 2010), and Canada (61; Spencer Stuart Board Index 2008). The average length of time that our respondents had served on their boards was 7.8 years, which is slightly over the average board tenure of directors in Spain and Canada (5 years; Spencer Stuart Board Index Canada and Spain; 7 years for Singapore). Most of our respondents had either Master’s or undergraduate-level education. Directors in our sample hailed from a variety of functional backgrounds, including marketing/sales, finance, research and development, engineering, operations, and law.

Survey Design

Due to the nature of directors’ responsibilities and the sensitive topics covered by our questionnaire, a low rate of response was a priori expected. Thus, several response facilitation approaches as well as N-Bias techniques were applied. By following Rogelberg and Stanton (2007); Randall and Fernandes (1991) and Podsakoff et al. (2003), several mechanisms were applied to prevent a substantive impact of a low response rate on the conclusions and simultaneously reduce the impact of common method bias: (i) the survey was conducted anonymously; (ii) a cover letter was included explaining the importance of the study for its implications on corporate governance in the international context; (iii) we used pre-notification and reminder notes several days before and after the first release; (iv) we used both mail and email copies to provide response opportunities with our respondents; and (v) survey feedback was promised to be delivered by email to all those directors who provide us with their email addresses.

Directors were asked to rate on a seven-point Likert scale from 1 (strongly disagree) to 7 (strongly agree) to agree or disagree with our statements. We carefully designed the questionnaire to prevent an excessive length (which reduces the likelihood of response) and to avoid item ambiguity (which will exacerbate common method biases). The different cultural characteristics of directors from each country were also taken into account when translating the questionnaire English into Spanish in order to ensure a correct and polite approach to the director. A pretest on several directors helped to achieve this aim.

Measures

Independent Variables

Each dimension of trustworthiness was measured with a scale developed by Mayer and Davis (1999) but adapted for the board context. Benevolence was measured with four items, including “The CEO is very concerned about my welfare”. The reliability for this measure was acceptable (α = .85). Integrity was measured with six items, including “I know the CEO will stick to his/her word,” and the scale reliability for this measure was high (α = .90). The ability measure consisted of seven items, including “The CEO is very capable of performing his/her job”. The reliability for this measure was high (α = .92).

Board performance evaluation practices were assessed with four items: (1) We conduct regular board evaluations; (2) We conduct evaluations of the board chair; (3) We evaluate fellow board members; (4) These evaluations are used frequently. The internal consistency of this measure was high (α = .94).

Dependent Variables

Resource provision was measured with five items developed by Westphal (1999) but adapted for board members. A representative item is “I act as a ‘sounding board’ to the CEO on strategic issues”. The reliability of this measure was reasonable (α = .74). Monitoring was measured with four items as developed by Westphal (1999) and adapted for our context; a representative item is “I monitor the CEO’s strategic decision making”. The reliability of this measure was also reasonable (α = .70).

Control Variables

Because participants hailed from different countries, it was important to control for the influence of cultural differences on the participants’ governance behaviors (i.e., monitoring and resource provision). We therefore used country dummies to identify cultural differences. Thus, we constructed two dummy variables C 1 and C 2, where C 1 is a dummy variable that takes the value of 1 for Canadian firms, and 0, otherwise; and C 2 is a dummy variable that takes the value of 1 for Singaporean firms, and 0, otherwise. We only incorporated two dummy variables to control for the three countries in order to prevent the well-know dummy tramp effect. We also controlled for whether the participant served as a board member on a for-profit or non-profit organization.

The estimated models are as shown in Eqs. 1 and 2.

$${\text{Monit}}_{i} = \beta_{0} + \beta_{1} {\text{In}}_{i} + \beta_{2} {\text{Ab}}_{i} + \beta_{3} {\text{Bnv}}_{i} + \beta_{4} {\text{BP}}_{i} + \beta_{5} {{\text{Pr}}_{i}} + \beta_{6} C_{1i} + \beta_{7} C_{2i} + \beta_{8} {\text {In}}*{\text{BP}}_{i} + \beta_{9} {\text{Ab}}*{\text{BP}}_{i} + \beta_{10} {\text{Bnv}}*{\text{BP}}_{i} + u_{it} $$
(1)
$$ {\text{RP}}_{i} = \beta_{0} + \beta_{1} {\text{In}}_{i} + \beta_{2} {\text{Ab}}_{i} + \beta_{3} {\text{Bnv}}_{i} + \beta_{4} {\text{BP}}_{i} + \beta_{5} {{\text{Pr}}_{i}} + \beta_{6} {\text{Ind}}_{i} + \beta_{7} {\text{Pwd}}_{i} + \beta_{8} {\text{In}}*{\text{BP}}_{i} + \beta_{9} {\text{Ab}}*{\text{BP}}_{i} + \beta_{10} {\text{Bnv}}*{\text{BP}}_{i} + u_{it} $$
(2)

where \( {\text{Monit}}_{i} \) stands for Monitoring, RP i stands for resource provision, In i stands for integrity, Ab i stands for ability, Bnv i stands for benevolence, BP i stands for board practices, Pr is a dummy variable that takes the value of 0 for non-profit firms, and 0 otherwise; and C 1 and C 2 stands for our country dummies.

Data Analysis

Descriptive statistics, correlations, and internal consistency reliability estimates for the measures are presented in Table 1. Figure 1 shows the histograms for dependent and independent variables.

Table 1 Descriptive statistics, correlations, and internal consistencies
Full size table
Fig. 1
figure 1

Histograms

Full size image

To determine the predictive power of the independent variables on the different governance behaviors, hierarchical multiple regressions were performed separately for monitoring and resource provision. These results are shown separately in Tables 2 and 3, respectively. In each hierarchical regression, control variables (for-profit organization and country dummies) were entered in the first step, and the primary-independent variables were entered in the second step. Interaction effects were tested by forming composite terms and assessing the incremental variance explained (Aiken and West 1991).

Table 2 Regression results for trustworthiness, board practices and monitoring: Model 1 (Eq. 1):\( {\text{Monit}}_{i} = \beta_{0} + \beta_{1} {\text{In}}_{i} + \beta_{2} {\text{Ab}}_{i} + \beta_{3} {\text{Bnv}}_{i} + \beta_{4} {\text{BP}}_{i} + \beta_{5} {{\text{Pr}}_{i}} + \beta_{6} C_{1i} + \beta_{7} C_{2i} + \beta_{8} {\text{In}}*{\text{BP}}_{i} + \beta_{9} {\text{Ab}}*{\text{BP}}_{i} + \beta_{10} {\text{Bnv}}*{\text{BP}}_{i} + u_{it} \)
Full size table
Table 3 Regression results for trustworthiness, board practices and resource provision: Model 2 (Eq. 2): \( {\text{RP}}_{i} = \beta_{0} + \beta_{1} {\text{In}}_{i} + \beta_{2} {\text{Ab}}_{i} + \beta_{3} {\text{Bnv}}_{i} + \beta_{4} {\text{BP}}_{i} + \beta_{5} {{\text{Pr}}_{i}} + \beta_{6} {\text{Ind}}_{i} + \beta_{7} {\text{Pwd}}_{i} + \beta_{8} {\text{In}}*{\text{BP}}_{i} + \beta_{9} {\text{Ab}}*{\text{BP}}_{i} + \beta_{10} {\text{Bnv}}*{\text{BP}}_{i} + u_{it} \)
Full size table

Results

Our hypotheses relating to the effects of ability on a director’s governance behaviors were not supported. Our first hypothesis (H1a), which suggests a negative relation between a director’s perceptions of the CEO’s ability and his or her monitoring activities, was not supported, since we did not find a positive relationship. Similar results were attained for Hypothesis 1b, which refers to ability and resource provision. Because of the unexpected results, we carefully examined the scatter plots to further investigate these results. As the scatter plots appeared to suggest the presence of a nonlinear relationship, we tested new versions of Models 1 and 2, where we included the square term of ability, considering the likelihood of a nonlinear relationship. We refer to these models, as Models 1b and 2b, and are shown in Eqs. 3 and 4, respectively.

$${\text{Monit}}_{i} = \beta_{0} + \beta_{1} {\text{In}}_{i} +\beta_{2} {\text{Ab}}_{i} + \beta_{3} {\text{Bnv}}_{i} + \beta_{4}{\text{BP}}_{i} + \beta_{5} {{\text{Pr}}_{i}} + \beta_{6} C_{1i} +\beta_{7} C_{2i} + \beta_{8} {\text{Ab}}_{i}^{2} + \beta_{9}{\text{In}}\,*{\text{BP}}_{i} + \beta_{10} {\text{Ab}}^{2}*{\text{BP}}_{i} + \beta_{11} {\text{Bnv}}*{\text{BP}}_{i} +u_{it} $$
(3)
$$ \begin{aligned} {\text{RP}}_{i} = \beta_{0} + \beta_{1}{\text{In}}_{i} + \beta_{2} {\text{Ab}}_{i} + \beta_{3}{\text{Bnv}}_{i} + \beta_{4} {\text{BP}}_{i} + \beta_{5}{{\text{Pr}}_{i}} + \beta_{6} C_{1i} + \beta_{7} C_{2i} + \hfill\\ \beta_{8} {\text{Ab}}_{it}^{2} + \beta_{9} {\text{In}_i}*{\text{BP}}_{i} + \beta_{10}{\text{Ab}_i}^{2} *BP_{i} + \beta_{11}{\text{Bnv}_i}*{\text{BP}}_{i} + u_{it} \hfill \\ \end{aligned} $$
(4)

where \( {\text{Ab}}_{{_{i}}}^{2} \) stands for squared ability, and the rest of the variables are defined as in Models 1 and 2.

According to the new estimations and regarding monitoring, we found a quadratic relation between a director’s perceptions of the CEO’s ability and his or her monitoring activities, as depicted in Fig. 2. The quadratic function has a break-point, which is a maximum, at 6.040, which indicates that the perception of values of ability under 6.040 will lead to higher levels of monitoring (β = 1.945, p < .001) but directors exercise less control for very high level of perceived CEO’s ability (β = −.161, p < .001), thus reducing monitoring, as proposed in Hypothesis 1a. Regarding resource provision, we also identified a quadratic relationship suggesting that a director’s lower perceptions of the CEO’s ability would motivate higher resource provision, but this behavior diminishes when the director’s perceptions of CEO ability are high (β = 1.811, p < .001; β = −.183, p < .001). The function also has a maximum, located in point 4.948, as shown in Fig. 3. Therefore, directors diminish their resource provision only for very high level of perceived CEO’s ability.

Fig. 2
figure 2

Quadratic relationship between monitoring & ability

Full size image
Fig. 3
figure 3

Quadratic relationship between resource provision & ability

Full size image

The hypotheses relating to the effects of CEO integrity on a director’s behaviors received mixed support. Hypothesis 2a, which proposes a negative relation between a director’s perceptions of the CEO’s integrity and his or her monitoring activities, was supported (β = −.763; p < .001). Hypothesis 2b, which predicts that a director’s perceptions of the CEO’s integrity would positively affect his or her resource provision activities, was supported in the first step of the model (β = .394; p < .01), although the effects are diluted in the last model (β = −.183; p > .05).

Regarding benevolence, only Hypothesis 3b was supported; we found that a director’s perceptions of the CEO’s benevolence positively affected his or her resource provision (β = .161; p < .01). However, hypothesis 3a was not supported; a director’s perceptions of the CEO’s benevolence did not seem to affect the extent of monitoring (β = .145; p > .05).

Our fourth hypothesis on the effects of board performance evaluation practices on board members’ resource provision and monitoring behaviors, received mixed support; board performance evaluation practices were positively related to directors’ resource provision (β = .258; p < .001), but not to monitoring (β = −.268; p > .05). The fifth hypothesis, which predicts moderating effects of board performance evaluation practices on the relation between a director’s perceptions of the CEO’s ability, benevolence, and integrity and his or her monitoring and resource provision behaviors, was also partially supported. The board performance evaluation practices interacted with the director’s perceptions of the CEO’s integrity when depicting monitoring (β = .148; p < .01), but other relations were non-significant either for monitoring or for resource provision. Finally, in order to analyze the impact of cultural differences on our results, we run Wald tests for Models 3 and 4, as shown in Tables 2 and 3. In both cases, we found that the inclusion of country dummy variables increases the explanatory power of both models, thus supporting the relevance of cultural differences on our analysis.

Discussion

Unlike much of the existing research on board governance that focuses on how to incent or motivate directors to perform effectively, we have examined directors’ behaviors, specifically, focusing on how directors’ perceptions of different facets of a CEO’s trustworthiness (i.e., ability, integrity, benevolence) affect their resource provision and monitoring behaviors. Our findings suggest that directors’ behaviors are influenced by their trustworthiness perceptions of the CEO and that different trustworthiness aspects have different mechanisms that impact directors’ behaviors. Our study has also considered how the existence of board performance evaluation practices can affect the directors’ governance, and how they can moderate the relations between the directors’ perceptions of the CEO’s trustworthiness and their monitoring and resource provision. Our results are highly generalizable; this research was conducted with an international sample (e.g., directors on boards in Canada, Spain, and Singapore), and we controlled for cultural differences (country dummies) that may affect board evaluation practices and typical governance behaviors.

Because CEOs and board members do not perform their duties in vacuum without any interpersonal relationships, we believe that our focus on the interpersonal dimensions of these relationships presents a realistic picture. In essence, it is too simplistic to state that CEOs should be “trustworthy”. By empirically examining how each element of trustworthiness affects directors’ governance behaviors, we are able to show the extent to which individual directors will monitor and provide resources. Interestingly, depending on the specific profile of the CEO’s trustworthiness, the director may monitor or provide resources, but not necessarily do both.

According to our findings, high levels of trustworthiness perceptions of CEO’s integrity lead to less monitoring by the director. Contrary to prior research (Creed and Miles 1996; Langfred 2004), we did not find a negative relationship between CEO’s ability and monitoring. Instead, the scatter plots suggested that a curvilinear effect may describe this relationship better. In fact, we detected a nonlinear relationship for monitoring, which indicates that lower levels of trustworthiness perceptions of ability lead to higher levels of monitoring and that monitoring decreases for a very high level of CEO’s ability. This finding is not consistent with the prior research on the impact of trust which shows that higher trust leads to lower monitoring (Langfred 2004) suggesting that trustworthiness perceptions and monitoring are not necessarily simple substitutes of each other in the board context. One possible interpretation of this result is that directors consider managerial monitoring as their primary duty and higher CEO’s ability motivates those directors to perform their monitoring task. As for resource provision, our results show that higher levels of trustworthiness perceptions of integrity and benevolence lead the director to provide more resources consistent with the prior trust research. Our study thus extends a growing body of literature which suggests that the assessment of someone’s ability is not the only important determinant of whether one will forge a reciprocal working relationship with a potential partner (Casciaro and Lobo 2008a, b). We also found a quadratic relationship between CEO’s ability and resource provision, which indicates that directors reduce their resource provisions when their perceptions of the CEO’s ability are high.

The results of our study also suggest that the antecedents of monitoring and resource provision are different; monitoring has a negative relationship with directors’ perceptions of CEO trustworthiness in terms of integrity, and resource provision is directly and positively related to the director’s perceptions of CEO benevolence and ability, but not with integrity. It is possible that directors’ perceptions of the importance of these behaviors may account for these differences. Resource provision may be seen as a personal favor to the CEO and hence social exchange incentives play a more important role, whereas monitoring may be seen as a less discretionary behavior and hence lower levels of CEO integrity which may cause damages to the firm invokes the director’s motivation to monitor more. Because monitoring and resource provision are both important to the effective functioning of the board and the organization, future research can examine additional antecedents of resource provision behaviors, with particular attention to any possible barriers.

We have also shown that while the board performance evaluation practices directly affect only resource provision by directors, these practices interact with a director’s perceptions of the CEO’s integrity to increase monitoring. Essentially, directors are more likely to provide resources to the CEO when the board members are evaluated frequently and effectively, and when the directors’ perceptions of the CEO’s integrity motivate them to reciprocate based on social exchange norms. However, we did not find any interaction effect with the CEO’s ability and benevolence. Further, those practices did not interact with the trustworthiness perceptions to enhance resource provision. These findings suggest that board evaluation practices may be an effective way to heighten directors’ awareness of their resource provision role when there is a social exchange relationship between the CEO and directors. Also, the director’s incentive to monitor the CEO and the effect of board performance evaluation practices are separate and each has an independent effect on the director’s behavior. Importantly, the mere presence of these board evaluation practices does not appear to influence the board members’ behaviors. Future research can examine if there are any specific board performance evaluation practices that are especially useful in this regard.

One of the important contributions of this study is that it is the first to examine the effects of trust on directors’ resource provision and monitoring. This study, therefore, presents an important step toward understanding the impact of interpersonal factors on how board members behave in the boardroom. Future research can build on this work by examining potential antecedents and moderators of this relationship. For example, social ties (Granovetter 1973; Westphal 1999) between a board member and the CEO may reduce the board member’s likelihood of monitoring, and they may cause the CEO to act more benevolently, thereby increasing the board member’s likelihood of providing more resources. Whereas this study has focused on the role of trust, it would be interesting to examine other important social exchange factors such as interpersonal justice (Masterson et al. 2000), especially from the CEO, in predicting board members’ behaviors.

While this study and many others have focused on unidirectional relationships between a trustor and trustee, trust perceptions between two parties are sometimes interdependent (Ferrin et al. 2007, 2008; Yakovleva et al. 2010). As such, it would be interesting to consider how a director’s governance behaviors may be interpreted by the CEO. By observing how a director monitors, (e.g., how the director processes information and the types of questions he or she asks), the CEO can assess the director’s ability. Increased monitoring may not necessarily lead the CEO to develop a positive perception of the director’s ability, because intense monitoring is not always necessary and can even be counterproductive. Intense monitoring, which signals low trust, may actually damage the relationship between the CEO and director by fostering negative feelings and suspicion (Cialdini 1996; McEvily et al. 2003). The CEO’s assessment of the director’s monitoring, then, would be based on both amount and the quality of such behaviors. A director’s monitoring can also provide opportunities for the CEO to observe and assess the director’s integrity, because his or her monitoring may reveal a set of principles that are being used to evaluate the CEO’s strategic plans and performance. If the CEO believes that the director monitors his or her performance in an honest, fair, and consistent manner, then the CEO is more likely to have a positive perception of the director’s integrity.

A director’s resource provision activities may also provide opportunities for the CEO to assess his or her ability, by observing the quantity and quality of resources are provided, which may signal the quality of the director’s human and social capital (Pfeffer and Salancik 1978; Hillman and Dalziel 2003). If the director gives useful professional advice or business ties with a powerful supplier, this may enhance the CEO’s perceptions of the director’s ability. A director’s resource provision activities may also signal his or her willingness to help the CEO, and therefore be interpreted as a proxy for the director’s benevolence. Further research can investigate these possibilities further.

Just as the director’s perceptions of the CEO’s trustworthiness affect the director’s governance behaviors, it would be interesting to examine how the CEO’s perceptions of the director’s trustworthiness affect the CEO’s behaviors. As noted by Ferrin et al. (2007, 2008), when one party holds positive perceptions of the other’s trustworthiness, he or she often behaves cooperatively, which in turn affects the other party’s perceptions. When the CEO has positive perceptions of a director’s trustworthiness, it would be interesting to examine if he or she then becomes more willing to work with the director by, for example, disclosing useful but sensitive information or by proactively seeking the director’s advice and resources. Because positive perceptions of integrity and benevolence indicate low risks of opportunism and non-reciprocity (Cruz et al. 2010; Ferrin et al. 2008), the CEO’s positive perceptions of a director’s trustworthiness may lead the CEO to exhibit trustworthy behaviors toward the director. Longitudinal research that captures how both CEO and directors’ perceptions of the other’s trustworthiness evolve over time is therefore warranted.

Despite the many contributions of this study, some limitations must be acknowledged. First, our study relied on self-report data. However, it should be noted that the board members’ perceptions of the CEO’s trustworthiness as well as their monitoring behaviors would be difficult for a third party to assess. Second, the sample in our study was relatively small. In consequence, we were precluded from using structural equation modeling, which would have taken into account correlations between our independent variables as well as measurement error. Third, our data was not longitudinal, which prevents us from making more definite claims about the causal relations between the variables in our hypothesized model. However, we should emphasize the difficulty in surveying active members of boards of directors; this population is limited in number, busy, and not readily incented to complete surveys, even for academic research.

Advocates of “good” corporate governance typically ground their arguments on agency theoretic logic, and call for a greater number of independent directors on the board. However, because there is limited research on the personal interactions between board members and CEOs, or within the board itself, we still do not know much about the effects of interpersonal dynamics on board governance (Hambrick et al. 2008). Insight into the interpersonal aspects of board member relationships provides us with a more nuanced understanding of the underlying forces that may affect the quality of the decisions being made. As such, it behooves us to continue to study the governance behaviors of corporate directors; the financial health and sound governance of businesses depends on it.