Master Thesis:
Regulatory Foci at the Top? How Directors’ and Chief Executive Officers’ Regulatory Foci Shape Governance Effectiveness
Word count (excluding references): 13,597
Abstract
This study addresses how directors’ motivational traits (i.e., regulatory focus) affect governance effectiveness (i.e., monitoring and advising) in isolation of and in conjunction with those of chief executive officers (CEOs). The results, derived from a unique multisource dataset that includes over 300 observations from Dutch directors and CEOs across 60 firms, indicate that a prevention focus in directors affects governance effectiveness more strongly and negatively in comparison to a promotion focus. At the interpersonal level (i.e., director–CEO), the findings are insignificant. However, the supplementary analysis reveals a congruence effect of the board’s collective regulatory focus and CEOs’ regulatory focus (i.e., board–CEO) on governance effectiveness. More specifically, the empirics show a positive prevention congruence effect on advising and a negative promotion congruence effect on advising. This study extends existing research on boards and provides unique micro-level insights into how directors and the dynamics between directors and CEOs shape governance effectiveness.
Introduction
Residing at the very top of organisations, non-executive directors (hereafter referred to as directors1) are a crucial part of firms’ governance mechanism (Zahra & Pearce, 1989). In line
with their fiduciary duties to company shareholders, these directors are primarily tasked with controlling and monitoring senior management’s actions (Fama & Jensen, 1983; John & Senbet, 1998), as well as advising and consulting with senior management on strategic actions (Hillman & Dalziel, 2003). In this context, directors provide not only a more objective perspective on strategy (Geletkanycz & Hambrick, 1997) but also access to external resources
1 This paper refers to non-executive directors, outside directors, or supervisory directors as directors, whereas
(i.e., networks to other organizations) (Hillman & Dalziel, 2003); they are likewise involved in hiring, firing, and deciding on senior management’s remuneration (Hillman & Dalziel, 2003; Pearce & Zahra, 1992; Veltrop, Molleman, Hooghiemstra, & van Ees, 2017; Westphal, 1999). Given the importance of senior management and directors, who reside at the upper echelons of organizations, strategy research has increasingly sought to improve the understanding of how the attributes of senior management and directors affect firms’ strategic decisions and thus shape organizational outcomes (e.g., Hayward & Hambrick, 1997; Hiller & Hambrick, 2005). In their seminal work on upper echelons (UE) theory, Hambrick and Mason (1984) highlight the significance of the psychological attributes of those residing at upper echelons. More specifically, UE theory states that the interpretations of strategic situations by those residing at upper echelons are based on their values, cognition, and personalities (Hambrick, 2007).
a proximal predictor of both performance and behavior (Johnson, Shull, & Wallace, 2011; Lanaj, Chang, & Johnson, 2012).
Whereas research regarding how CEOs’ regulatory focus influences organisational outcomes and strategic decisions has gained some traction recently (e.g., Bilgili et al., 2018; Gamache et al., 2015; Wallace, Little, Hill, & Ridge, 2010), research on how directors’ regulatory focus affect their monitoring and advisory work in boards remains lacking. This is a crucial omission because a regulatory focus highlights preferences for strategic action, namely eagerness (promotion focus) and vigilance (prevention focus) (Gamache et al., 2015), both of which can influence goal striving (Lanaj, Chang, & Johnson, 2012) and directors’ engagement in monitoring and advising. In this sense, diverse regulatory foci may potentially have different effects on monitoring and advising. More specifically, whereas directors’ prevention focus may enhance their engagement in monitoring, directors’ promotion focus may enhance their engagement in top management advising.
in monitoring and advising, this study also acknowledges how the director–CEO regulatory fit may influence directors’ engagement in monitoring and advising.
This research aims to address this gap and hypothesizes that directors’ prevention focus is positively associated with monitoring and negatively associated with advising. Contrastingly, I anticipate that the effects are reverse for directors’ promotion focus. To depict a more complete picture, I additionally hypothesize that a prevention-focused director–CEO
regulatory fit optimizes monitoring, whereas a promotion-focused director–CEO regulatory fit
optimizes advising. The reasoning is primarily based on creating a mutual understanding of the necessity and desire to perform such tasks, as well as on establishing improved communication and motivation (Stam, van Knippenberg, & Wisse, 2010; Venus, Stam, & van Knippenberg, 2013; Wexley, Alexander, Greenawalt, & Couch, 1980). A sample of 61 Dutch firms and 302 responses from directors who work in the housing, health, and education industries was used to empirically test the hypotheses.
By empirically examining how directors’ regulatory focus relates to effective monitoring and advising, this study makes several contributions to existing research. First, it expands current studies on the influence of leaders’ regulatory focus. These have primarily concentrated on examining the effects of low/middle management’s regulatory focus on individual-level outcomes (e.g., Kark & Van Dijk, 2007; Neubert et al., 2008; Shin et al., 2017) and the impact of CEOs’ regulatory focus on firm level-outcomes (e.g., Gamache et al., 2015) by investigating how directors’ regulatory focus influences governance effectiveness (i.e., monitoring and advising).
Additionally, the current research has revealed that directors’ prevention focus is negatively related to advising, whereas a promotion focus does not significantly affect either monitoring or advising. Overall, these findings indicate that directors’ prevention focus is more influential on governance effectiveness.
The third contribution of this paper relates to the notion of interpersonal regulatory fit. Although the findings indicate that at the director–CEO level, the effects of both prevention and promotion congruence on either monitoring or advising are insignificant, the supplementary analysis has revealed that at the board–CEO level, both prevention and promotion congruence significantly affect board advising. Yet, surprisingly, the effects suggest that a prevention fit leads to improved board advising, whereas a promotion fit hampers it. Interestingly, the results indicate that the prevention congruence effect is primarily driven by the board, whereas the promotion focus congruence effect is primarily driven by CEOs. This indicates the difficulty of advising highly promotion-focused CEOs. Additionally, these findings suggest that directors consider the regulatory focus of CEOs and provide less advice to strongly promotion-focused CEOs.
Overall, this research contributes to the understanding of the micro-level level antecedents to governance effectiveness. Doing so is essential for the literature, as firm-level outcomes are not only driven by but also cannot be truly understood without considering individuals and their interactions, which mainly comprise firms (Felin & Foss, 2005). More specifically, this research explores how directors’ psychological traits, such as their regulatory focus, influence board effectiveness (in terms of monitoring and advising), which improves the understanding of when and why acquisitions and governance failures occur.
followed by a description of the data, the data analysis, methods, empirical results, and the supplementary analysis. The research concludes with a discussion of the results, followed by an acknowledgement of limitations and the agenda for future research.
Theoretical background
Upper echelons theory
Corporate boards have attracted a variety of research on them, with UE theory being one of the frequently applied lenses (Johnson et al., 2017). UE theory is based on bounded rationality (Simon, 1957). Bounded rationality, in turn, influences the environmental clues that are recognized, how these clues are interpreted, and the actions taken in response to such clues (Hambrick & Mason, 1984). This is especially relevant, as the upper echelons of organizations receive more complex and ambiguous information than they can cope with. Consequently, those at upper echelons use their previous experiences and preferences to deal with this information overload (Cho & Hambrick, 2006). Specifically, UE theory argues that those at the top of an organization act based on their personal interpretations of situations. These interpretations are a function of executives’ values and personalities (Hambrick, 2007). Accordingly, at its core, UE theory is a theory of information processing. Given the information overload that those at upper echelons deal with, they tend to fall back on their cognitive base (their values, personality, and experiences) (Cho & Hambrick, 2006). Essentially, UE argues that organizations are a reflection of their top management (Carpenter, Geletkancz, & Sanders, 2004).
proxies and measure the underlying psychological attributes (e.g., Carpenter, Geletkancz, & Sanders, 2004). In response to these calls, governance research has recently documented the crucial role of upper echelons’ regulatory focus on firms’ strategic actions (Chen, Meyer-Doyle, & Shi, 2018).
Regulatory focus
Regulatory focus theory is a fundamental motivational theory (Scheepers, Ellemers, & Sassenberg, 2013) based on the hedonic principal (Higgins, 1997), which deals with how individuals avoid pain and approach pleasure. In this sense, regulatory focus theory is concerned with self-regulation, which influences processes and motivations that are engaged in regulating affect, behavior, and cognition when pursuing goals (Carver & Scheier, 1998). Regulatory focus theory also argues that individuals pursue their goals via either a promotion focus or a prevention focus (Higgins, 1997). Importantly, promotion and prevention foci represent orthogonal constructs rather than the opposite sides of the same construct (Förster, Higgins, & Bianco, 2003). This fact implies that individuals can score high on both promotion and prevention foci. Although differences between promotion- and prevention-focused individuals may result in variations in performance and decision making (Higgins, 1997), both promotion- and prevention-focused individuals can engage in successful goal pursuit (Lanaj, Chang, & Johnson, 2012).
success for strongly promotion-focused individuals, by the same token, it increases the chances of risk-taking for individuals (Higgins, 1997).
By contrast, individuals who have a strong prevention focus are more sensitive to negative stimuli (Crowe & Higgins, 1997). Prevention-focused individuals view goals as an opportunity to meet their responsibility and are driven by ensuring security, criticism, and the fear of failure. Accordingly, they tend to take on less risks and make fewer mistakes. Prevention-focused individuals therefore reduce the possibility of incurring losses, which, in turn, minimizes risk.
Over the last decades, regulatory focus theory has attracted much research attention from a wide range of disciplines, including, for example, psychology (e.g., Higgins, 1997), decision making (e.g., Crowe & Higgins, 1997), leadership (Kark & Van Dijk, 2007), and the organization and corporate governance contexts (e.g., Gamache et al., 2015). In the context of UE theory, a regulatory focus can affect the intentions of those at the top of organizations and how they leverage their skills when they take actions (Chen et al., 2018), ultimately affecting firm-level outcomes. Although few researchers, thus far, have applied a regulatory focus to the context of upper echelons, a regulatory focus offers interesting insights from a governance perspective because it highlights preferences for strategic action, namely eagerness (promotion focus) and vigilance (prevention focus) (Gamache et al., 2015), both of which affect goal striving (Lanaj, Chang, & Johnson, 2012).
advantage is that an individual’s chronic regulatory focus is rather stable over time (Higgins, 1997; Higgins et al., 2001; Strauman, 1996).
Research has long understood the strategic implications of regulatory focus theory. For instance, Förster, Higgins, and Bianco (2003) have demonstrated in a series of experiments that a promotion focus framing leads to increased performance speed and decreased accuracy. On the other hand, Förster, Higgins, and Bianco (2003) have shown that individuals perform better in error detection tasks when they are in a prevention focus framing. More recently, scholars have begun to explore the strategic implications of regulatory focus theory in the organizational and governance contexts. For example, Kark (2007) integrates motivation theory and leadership theory in his paper while simultaneously shedding light on the effect of regulatory focus and monitoring. Specifically, Kark (2007) argues that the two regulatory foci are basically two opposing motivations. While a prevention focus motivates individuals to strive for stability, a promotion focus motivates them to strive for change and improvement. Thus, a prevention focus drives individuals to ensure safety and maintain the status quo, whereas a promotion focus encourages individuals to grab the advantages of new, out-of-the box behaviors. In this context, Kark also suggests that prevention-focused leaders are likely to be associated with a monitoring leadership style.
Chung and Tsai (2009) have highlighted that promotion-focused individuals guard against the error of omission, whereas prevention-focused individuals guard against errors of commission. Additionally, Wallace, Little, Hill, and Ridge (2010) find that a strong prevention focus has diverse effects on top managers’ and CEOs’ performance ratings—whereas the effect on performance rating is negative for top managers, the effect for CEOs is insignificant.
Finally, Gamache et al. (2015) have discovered that promotion-focused CEOs engage in more mergers and acquisitions than do prevention-focused CEOs. In this context, Gamache et al. (2015) have also highlighted that boards should consider CEOs’ regulatory focus when deciding on how to direct CEOs. For example, they have proposed that promotion-focused CEOs need more oversight, whereas prevention-focused CEOs might need more advice and encouragement to take on risky ventures.
Despite this crucial function of the board, especially of directors, little is known about how their chronic regulatory focus influences their engagement in monitoring and advising CEOs. In line with the argument of Brockner et al. (2004), I argue that the different regulatory foci are beneficial for specific governance tasks.
Director regulatory focus and monitoring
at the frontline of defense in protecting shareholders. All in all, monitoring effectiveness decreases the possibility that CEOs who are not well suited for the position are employed or retained and that ill-fitting actions go unnoticed by the board (Hambrick, Misangyi, & Park, 2015).
However, obstacles, such as information asymmetry, unfavorable group dynamics, or external job demands, can have a negative influence on monitoring activities (Boivie et al., 2016). These obstacles to monitoring decrease the probability that monitoring actions are performed well. Accordingly, individuals taking on the monitoring role of CEOs need to be well suited to perform the task. Previous research has highlighted that prevention-focused individuals are more likely to make careful, systematic decisions (Fredrickson & Mitchell, 1984) and to conduct more effective due diligence (Brockner et al., 2004). The literature also suggests that prevention-focused leaders set up clear rules for followers, check details, correct misbehavior, and encourage compliance with these rules (Hamstra, Sassenberg, Van Yperen, & Wisse, 2014). Taken together, the mechanisms discovered through previous research suggest that prevention-focused directors, on average, perform especially well in a monitoring role and are more likely to be more active in such roles, as these types of tasks are natural to them. Therefore, I hypothesize the following:
H1: A positive relationship exists between directors’ prevention focus and directors’
engagement in monitoring.
monitoring activities, one can infer that a promotion-focused director might find it more difficult to exercise the vigilance and detail orientation required in monitoring CEO decision making. Besides, given the naturally higher risk acceptance of promotion-focused directors, they might start directing and governing CEOs’ behavior at a later point in time. Overall, this suggests that promotion-focused directors are less likely to closely monitor CEOs’ decision making. Basing on this, I hypothesize the following:
H2: A negative relationship exists between directors’ promotion focus and directors’
engagement in monitoring.
Regulatory focus and advising
H3: A positive relationship exists between directors’ promotion focus and directors’
engagement in advising.
Individuals with a strong prevention focus can be beneficial in some stages of a firm’s development because they help avoid losses, which can be an advantage when organizational resources are very limited (Fitzsimmons & Douglas, 2011). In the context of the advice given by directors to CEOs, there are circumstances, such as some sort of crisis, in which prevention-focused advice is appreciated. Thus, one can infer that in times of crisis advising, the engagement of prevention-focused directors increases.
In the absence of a crisis or another situation that demands a firm to increase vigilance, the advising task can prove to be more challenging for directors with a strong prevention focus. There are several reasons for this. First, previous research has shown that prevention-focused individuals have a strong desire to maintain the status quo (Kark, 2007). This desire results in less-innovative or less-creative ideas to be brought forward by directors. Second, prevention-focused individuals have been previously linked to a reduced exploration of alternatives (Higgins, 1997; Crowe & Higgins,1997). Given that idea generation and coming up with alternative courses of actions are inherent elements of advising, I hypothesize the following:
H4: A negative relationship exists between directors’ prevention focus and directors’
engagement in advising.
Regulatory fit
isolation from other people (Righetti, Finkenauer, & Rusbult, 2011). Specifically, people often interact with other people when they are pursuing goals, and the people with whom they interact potentially facilitate or hamper their goal pursuit (Kelley, 1983).
An interpersonal regulatory fit describes a situation in which the regulatory focus of people who are interacting—to pursue a goal—is the same. Therefore, an interpersonal regulatory fit and task regulatory fit have similar consequences, including motivational benefits (Johnson et al., 2017). Additionally, an interpersonal regulatory fit facilitates communication and understanding between the different parties involved (Stam, van Knippenberg, & Wisse, 2010; Venus, Stam, & van Knippenberg, 2013; Wexley, Alexander, Greenawalt, & Couch, 1980) and increases interpersonal comfort, compatibility, and work coordination (Kark, Van Dijk, & Esformes, 2011). Hamstra, Sassenberg, Van Yperen, and Wisse (2014) have studied the circumstances under which followers feel valued by their leader and found a positive effect of a regulatory fit. In a similar vein, Hamstra et al. (2013) have found that a regulatory fit leads to more favorable performance evaluations.
CFOs affects the growth of firms’ scope. Overall, a regulatory fit and especially interpersonal regulatory fit theory (Righetti, Finkenauer, & Rusbult, 2011) have complemented the emerging literature dealing with employee–supervisor fit (Johnson et al., 2015).
Regulatory fit and monitoring
In the context of monitoring a prevention fit between directors and CEOs, high engagement in monitoring may likely occur. Specifically, a prevention fit has several key advantages in the context of monitoring engagement. First, as highlighted previously, a prevention-focused director is naturally more vigilant (Gamache et al., 2015) and performs better in error detection tasks (Förster, Higgins, & Bianco, 2003). Second, as highly prevention-focused CEOs share such characteristics with directors, these CEOs are more likely to see the necessity of monitoring and thus will provide the board with the information it needs.
Another reason why a prevention fit is beneficial for monitoring engagement is that the congruence effect results in advantages, such as facilitating communication and mutual understanding (e.g., Stam et al., 2010), as well as increasing comfort, compatibility, and work coordination (Kark, Van Dijk, & Esformes, 2011). As a result, social barriers to monitoring are more easily overcome, as both parties perceive monitoring as a necessity instead of a negative and mistrusting action.
H5: The interpersonal regulatory (i.e., prevention) congruence effect between directors and
CEOs is positively related to engagement in monitoring.
Regulatory fit and advising
message/vision (Stam et al., 2010). When transferring this information to the context of advising, one can assume that the advice becomes more effective when a regulatory fit exists between the individuals involved. Moreover, a promotion regulatory fit is, in general, perceived to enhance creativity (Wu, McMullen, Neubert, & Yi, 2008), suggesting that the uniqueness of advising increases further when directors experience a regulatory fit with CEOs.
The first experimental evidence for this behavior has been brought forward by Righetti, Finkenauer, and Rusbult (2011), who have suggested that promotion-focused individuals seek advice from others and are also more open to receiving advice. Additionally, they have found that promotion-focused individuals are more likely to recognize the presence of a regulatory fit, which strengthens the motivational benefits of an interpersonal fit. In the context of this study, it can be argued that promotion-focused directors are more eager to find new opportunities and are less likely to overlook essential developments that can yield improvements in advising. As CEOs share promotion focus features with directors, it becomes more likely that CEOs are not only open to advice but also explicitly seek out advice from directors so that the dyad engages in a continuous exchange of ideas, opportunities, and advice. Therefore, I propose the following:
H6: The interpersonal regulatory (i.e., promotion) congruence effect between directors and
CEOs is positively related to engagement in advising.
Methodology
Research setting
has been primarily used by Dutch non-profit organizations in the education, health, or housing sector.
Dutch law requires the use of a two-tier board system, thus formally separating the board into the management board, composed by the CEO and the remaining executives, and the supervisory board, composed by directors. Although the one- and two-tier systems have fundamental differences, the tasks and responsibilities of supervisory directors in a two-tier system are largely equivalent to those of non-excusive or outside directors in a one-tier system (Hooghiemstra & Van Manen, 2004), as they are separate from management and are not fulltime members of the firm. This research setting is particularly well suited for the analysis, as using non-profit organizations instead of stock-listed companies is more likely to result in unbiased responses. This is because any publicity concerning the functioning of stock-listed boards might affect share price (Veltrop, Molleman, Hooghiemstra, & van Ees, 2018).
Sample and Procedure
One of the most challenging aspects of understanding the dynamics in boards of directors is accessing directors’ complete responses (Westphal & Stern, 2007). In ensuring access to boards, the web-based tool aims to create a win-win situation by providing organizations with feedback on the functionality of boards, which can be used as input for the yearly self-evaluation.
those of supervisory boards, received unique access codes, which they used to securely log into the website.
The multisource survey contains responses from more than 450 Dutch executives and directors from 66 Dutch firms. After incomplete responses were deleted, the sample contained 396 (88%) responses from 63 (93%) firms. Additionally, one firm had to be excluded because the CEO did not complete the survey sufficiently. Importantly, this did not significantly change the population of our sample. Specifically, neither the F-test nor the Kolmogorov-Smirnov test showed a significant difference in terms of the directors’ demographics. The average age of the respondents is 56.72 years (sd=8.6), and their average tenure as a member of the board of directors is 4.1 years (sd=4.35). The respondents are primarily male (65.7%). Moreover, 22% of the participants belong to the management board, and 78% belong to the supervisory board. The average size of the management board is 1.59 (sd=0.905), and the average size of the supervisory board is 5.32 (sd=1.022). Most participating organizations operate in the health sector (39.3%), followed by the housing sector (35.1%). Another 11.1% of the firms in the sample are in the education sector. The remaining firms belong to other industries.
After the CEO variables were created, the CEO responses were dropped to avoid comparing the CEOs with themselves and creating false congruence effects. Instead, the relationship between directors and CEOs was analyzed. Accordingly, our sample was reduced to 335 observations. Only the relationship between directors and CEOs is of interest, so the responses of other executive directors were also excluded, resulting in a final sample of 302 observations.
Common method variance
Organ, 1986). Bias is one of the primary sources of systematic measurement error. If one does not account for CMV, the results might be biased. More precisely, the data might not reveal anything about the theoretical relationship being studied. Instead, it will shed light on the “artificial cognitive maps of reality hidden in the respondents’ minds” (Chang & Eden, 2010). Several measures were taken to ensure that CMV is not an issue in this analysis. In line with the suggestions of Podsakoff et al. (2003), the conceptual model includes not only Likert scale measures but also further objective measures, such as demographics and tenure, to decrease CMV.
Most importantly, however, the conceptual model combines individual, interpersonal, and peer-rated variables, as well as self-rating. Accordingly, the model largely avoids using the same source. The measures were also selected from different sections of the survey, creating distance among the measures. Harman’s single-factor test was performed to ensure that the measures helped mitigate the CMV effect. Specifically, the one-factor test was applied to check if a single latent factor could account for all the constructs. Commonly, Harman’s one-factor test assumes that CMV is present when only a single factor appears from factor analysis or when the first factor accounts for more than 50% of the variance among variables (Podsakoff & Organ, 1986). For the models tested in this study, the explained variance is 10.43%. Accordingly, CMV is unlikely to inflate or deflate the observed estimates of the relationships. In short, although the data stem from a cross-sectional survey, it is unlikely that CMV affects the results because of the different data sources used in the conceptual model.
Measures
Engagement in monitoring. Director engagement in monitoring was measured on a
[name] effective in monitoring top management?” and “To what extent does [name] monitor top management strategic decision making?” The Cronbach’s alpha of these items is .7330.
Engagement in advising. Advising is also measured on a seven-point Likert scale ranging
from (1) very little to (7) a very great deal, with the CEO accessing each director via two items adapted from the work of Carpenter and Westphal (2001). The items were “All in all, to what extent is [name] effective in advising top management?” and “To what extent does [name] provide top management with advice on strategic issues?” The Cronbach’s alpha of these items is .8766.
Regulatory focus. The regulatory focus of the directors and the CEO, in turn, is accessed
through 12 items on a seven-point Likert scale. Six of these items reflect a prevention orientation (e.g., “In general, I am focused on preventing negative events in my work”) and have a Cronbach’s alpha of .6012. The remaining six reflect a promotion focus (e.g., “I frequently imagine how I will achieve my hopes and aspirations in my work”) and have a Cronbach’s alpha of .7263. This scale was adapted from that by Lockwood, Jordan, and Kunda (2002) to fit the board context. As the Cronbach’s alpha of prevention focus was below the recommended cutoff of .7 (Mackenzie, Podsakoff, & Podsakoff, 2011), an additional analysis was performed to investigate the causes of the low Cronbach’s alpha. This analysis revealed that one item, “I frequently think about how I can prevent failures in my work,” loaded on a different construct. Therefore, I excluded this item from the analysis, which improved the alpha to 0.6905.
Interpersonal regulatory fit. Congruence (i.e., fit, similarity, or agreement) between two
Although other methods (e.g., difference scores and moderation effects) can also yield interesting insights into the effect of fit, they are frequently critiqued because, for example, difference scores are related to several potential problems. These problems include confounding effects, conceptual ambiguity, and the assumption that positive and negative differences have the same effect (Edwards, 1993). Polynomial regressions overcome the shortcomings of other methods. Specifically, polynomial regression is a statistical approach that examines the extent to which combinations of two predictor variables—in this case, director regulatory focus and CEO regulatory focus—relate to an outcome variable, in this case, director monitoring and advising. In contrast to other approaches, such as difference scores, polynomial regression provides more specific information about the joint effect of director and CEO regulatory focus on the outcome variable, as well as about the congruence between director and CEO regulatory focus. More specifically,
(2) Z1,2 = b0 + b1X + b2Y + b3X2 + b4XY + b5Y2 + bC + e,
where Z1 and Z2 represent director monitoring and advising, respectively, X and Y represent CEO regulatory focus and director regulatory focus, respectively, and C represents a vector of the control variables (e.g., tenure, gender, board size).
Controls. This research controls for a variety of board-, industry-, and individual-level
Besides, a wide variety of personal and professional characteristics of board members and CEOs (for Hypotheses 5 and 6) have been included, as they may influence engagement in monitoring (Veltrop, Molleman, Hooghiemstra, & van Ees, 2018) and advising. Information on both the age and tenure of CEOs and directors were coded from annual reports, in combination with information from the Dutch Chamber of Commerce. Age and tenure (Hypotheses 5 and 6) were included because they can both affect engagement in monitoring and advising.
Additionally, I controlled for the gender (“1” denotes male and “2” denotes female) of directors because, on the one hand, it has been well established that the performance ratings of females are worse than those of their male colleagues (Lyness & Heilman, 2006); on the other hand, it has been suspected that female directors perform the monitoring function better than male directors do (García Lara, García Osma, Mora, & Scapin, 2017). CEOs’ gender was also included.
Finally, I used industry dummies to control for industry effects, as specific industries might require a higher level of monitoring and advising. As fixed effects are used to test the first four hypotheses, industry effects are only required to test the remaining hypotheses.
Data analyses
This study uses multilevel analysis to test the hypotheses (cf. Koopmann et al., 2016; Pollack, Vanepps, & Hayes, 2012; Yu & Zellmer-Bruhn, 2018). Directors (level 1) were nested within the 61 different organizations (level 2). To test the first four hypotheses, I used firm fixed effects while also applying random effects for the two latter hypotheses, as including fixed effects would account for all CEO-level variance.
As previously mentioned, I used polynomial regression with response surface analysis to test Hypotheses 5 and 6. Polynomial regression with surface response analysis highlights the degree to which each independent variable influences the outcome variable. This has the advantage of providing a much more detailed view of the congruence or incongruence effects of two variables (Byza et al., 2017). Following the recommendation of Edwards (1994), I mean centered the polynomial terms to ease the interpretation response surface indicators. The polynomial regression is based on the following formula:
1) Z1,2 = b0 + b1X + b2Y + b3X2 + b4XY + b5Y2 + bbC + e.
To gain an initial understanding of whether polynomial regression is suitable for the analyses, I conducted an F-test to determine whether the change in R-square between the control model (including the main effects X and Y) and the polynomial regression model (including the second-order terms, i.e., X2, XY, and Y2) is significant. When the change in R-square is significant, a non-linear relationship (Atwater, Ostroff, Yammarino, & Fleenor, 1998) and the appropriateness of the polynomial regression are confirmed.
indicators a1 and a2 represent the slope and curvature of the congruence line, respectively. The indicator a1 is calculated as the sum of b1 and b2, whereas the indicator a2 is the sum of b3, b4, and b5. Concerning incongruence, lines a3 and a4 represent the slope and the curvature, respectively. The slope a3 is calculated by subtracting b2 from b1. Finally, the incongruence curvature line a4 is calculated by subtracting b4 from b3 and subsequently adding b5.
Before conducting the analysis, I screened each analysis for influential cases by using Cook’s distance statistics. The results indicate one influential case. Upon further inspection, it became evident that there was an error in the data, which confused the age and the tenure. I corrected this issue, and Cook’s D statistics no longer revealed any influential cases.
Validity tests
I performed a principal component analysis of all measures in the survey. The results returned one eigenvalue greater than one for each component, thus indicating the appropriateness of combining the items into single factors. Furthermore, each Cronbach’s alpha is greater than the threshold of 0.6 (Bagozzi & Yi, 1988). All in all, these results indicate good inter-item reliability. Although the correlation table suggests no evidence of multicollinearity in the models, I additionally computed the variance inflation factor (VIF). The results show that the highest VIF is 3.45, whereas the lowest one is 1.15 for the analysis of the effect of a prevention fit on monitoring. The results for the effect of a promotion fit on advising are rather similar; the highest VIF score is 3.092 and the lowest is 1.13. These values are below the suggested cut-off value of 10 and greater than 1 (Chatterjee, Hadi, & Price, 2000). Therefore, one can confidently rule out multicollinearity for the models.
Results
prevention focus (r=−.113; p=.045), directors’ promotion focus (r=−.122; p=.035), and monitoring. These correlations are largely in line with the predictions made in that they show that the added control variables are significantly important to avoid omitted variable biases. Moreover, the correlations support the notion that directors with a high promotion focus perform less engagement in monitoring tasks.
Surprisingly, one can observe a strong correlation between monitoring and advising (r=.746; p=.001). This issue might come from the fact that both measures are rated by CEOs and address the same person. This correlation seems very high at first glance. However, the shared variance between monitoring and advising is .55. This result suggests that individuals perform similarly in both monitoring and advising functions, but, at the same time, a substantial amount of variance exists between monitoring and advising. Accordingly, they can be regarded as unique constructs. Similarly, CEOs’ prevention and promotion foci are significantly correlated (.372; p=.000); this casts some doubt in whether the two items are, in fact, separate constructs. However, as the two items only share .138 variance, they certainly represent independent constructs. Interestingly, no significant correlation between directors’ promotion and prevention foci has been found (.100; p=.08).
--- INSERT TABLE 1 ABOUT HERE
---
Hypothesis testing
their engagement in monitoring in the board. Although a significant relationship was found (p=.021), the effect indicates a negative relationship (β=−.1632). Accordingly, Hypothesis 1 is not supported. Hypothesis 2 argues that a negative relationship exists between directors’ promotion focus and monitoring. Model 3, however, shows that a promotion focus does not significantly influence monitoring (β=−.0319; p=.532). Accordingly, Hypothesis 2 is not supported. Model 4 in Table 3 shows the control model for the hypothesis assessing engagement in advising. Model 5 tests Hypothesis 4, which assumes a negative relationship between directors’ prevention focus and advising. Indeed, Model 5 shows a significant negative relationship (β=−.1175; p=.036), thus supporting Hypothesis 4. Model 6 tests the third main effect hypothesis, which argues a positive relationship between directors’ promotion focus and advising. However, the third hypothesis is not supported (β=−.02489; p=.697).
--- INSERT TABLE 2 AND 3 ABOUT HERE
---
Tables 4 and 5 present the results of the polynomial regression, and Figures 1a to 1b show the corresponding surface plots. Table 4 tests the polynomial regressions on monitoring and includes Models 7 and 8, in which Model 7 represents the control variables and the main effects (i.e., X and Y) of the polynomial regression. Model 8 adds the second-order polynomial terms (i.e., X2, XY, and Y2). Table 5 tests the polynomial regressions on advising and includes Models 9 and 10, in which Model 9 represents the control variables and the main effects of the polynomial regression and Model 10 adds the second-order polynomial terms.
--- INSERT TABLE 4 ABOUT HERE
Model 8 tests Hypothesis 5, which predicts a congruence effect between directors and CEOs, which, in turn, is positively related to engagement in monitoring. In line with the suggestion of Edwards (1994), I first analyzed if the second-order polynomial terms (i.e.,, X2, XY, and Y2) together significantly improve the variance explained in the monitoring variable. In comparison to Model 7, which only includes the control variables and the main effects (i.e., X and Y), Model 8 does not significantly improve the R-square (ΔR2=.0058; p=.631), thus
casting some doubt on the validity of Hypothesis 5. Nonetheless, I continued to analyze the response surface to gain a full understanding of the congruence effect. To support Hypothesis 5, one would expect a positive and significant congruence slope coefficient (i.e., a1). However,
Model 8 shows that the congruence slope coefficient (i.e., a1) is insignificant (p=.08) and
negative (−.18), so the results do not support the hypothesis.
Hypothesis 6 states that the congruence between directors and CEOs positively affects engagement in advising. Accordingly, to find support for this hypothesis, one would expect to find a positive significant slope coefficient of the congruence line. However, the change in R-square is insignificant (R2=.0221; p=.0899), and so is the congruence slope coefficient (a1=−1.44; p=.445). Therefore, the results of the analysis do not support Hypothesis 8.
--- INSERT TABLE 5 ABOUT HERE
---
Figures 1a and 1b show the response surface plots of prevention focus congruence on monitoring and promotion focus congruence on advice. In both figures, the congruence line (X=Y) corresponds to the line on the graph’s floor, which begins on the far left corner and proceeds to the near right corner, and the incongruence line (X=−Y) corresponds to the line on the graph’s floor and ranges from the far right corner to the near left corner.
INSERT Figures 1a and 1b ABOUT HERE ---
Supplementary analysis
In response to the counterintuitive (Hypothesis 1) and the insignificant results (Hypotheses 2 and 3), supplementary analyses were carried out to further explore the findings. The method used for this is similar to that described by Gamache et al. (2015).
First, the aim was to better understand why the relationship between prevention focus and monitoring was negative and not positive, as predicted. Dummy variables were therefore created for directors with high and low prevention foci, in which the high prevention focus dummy indicated an above-average prevention focus and the low prevention focus dummy indicated a below average prevention focus. Additionally, two more dummy variables were created—one denotes a very low prevention focus when the director scored among the bottom 25%, and the other denotes a very high prevention focus when the director scored among the top 25%. These variables were included separately in the model to test Hypothesis 1 (see Table 6). The hypothesis testing was conducted in line with the methodology described previously. The results of the analyses indicate that monitoring engagement is perceived worse (β=−.250;
p=.005) for high prevention-focused individuals, thus supporting previous results. However,
the analysis also shows that the effect of a low prevention focus has, in fact, a significant positive effect on engagement in monitoring (β=.250; p=.005). For very high levels of prevention focus, the results are insignificant (β=−.044; p=.664), whereas for very low levels of prevention focus, the effect is significant and positive (β=.272; p=.008). Although the results of the supplementary analysis remain counterintuitive, the robustness of the findings has been strengthened. The potential reasons for the counterintuitive results are discussed in the next section of the thesis.
INSERT TABLE 6 ABOUT HERE ---
Second, the research aims to better understand why no negative effect exists between a promotion focus and monitoring (Hypothesis 2). In line with the methodology used above, four dummy variables were created, and they denoted four levels of promotion focus: low and high promotion foci and very low and very high promotion foci. Subsequently, the dummies were separately included in our model to test Hypothesis 2 with the use of the methodology described previously. At high levels (i.e., above average) of promotion focus, the results become highly negative and significant (β=−.244; p=.006), conclusively supporting the hypothesis at high levels of promotion focus.
Third, the results of Hypothesis 3 were further investigated in Table 7 by applying the previously described procedure, which includes creating four dummy variables denoting low and high promotion foci, as well as very low and very high promotion foci. The cutoff values are the same, as described above. Including the dummies in the model to test Hypothesis 4 revealed the following results: at very high levels (β=.070; p=.483), at very low levels (β=.110;
p=.316), and at low levels (β=.001; p=.991) of promotion focus, the relationship between a
promotion focus and advising remains insignificant. However, at a high level of promotion focus, the results are significant and negative (β=−.220; p=.047). In contrast to the hypothesis, which predicts a positive significant relationship between promotion-focused directors and engagement in advising, the supplementary analysis has revealed that the relationship with advising is negative for high levels of promotion focus.
--- INSERT TABLE 7 ABOUT HERE
Another approach that was explored to better understand the counterintuitive results of our main effect hypotheses was to identify the effect of gender in the data (see Table 8). Specifically, the data continuously found that female directors are more negatively rated in engagement in advising (see Models 4 and 11; β=−.33; p<.001) than their male counterparts are, suggesting that our results could suffer from a bias against female directors (about 30% of the sample is made up of female directors). This is especially relevant in the current study, as previous research has found a significant gender effect on regulatory focus, so women are more strongly associated with a prevention focus, whereas men are more strongly associated with a promotion focus (McKay-Nesbitt, Bhatnagar, & Smith, 2013). Yet, in this sample, such an association could not be found (p=.710 for prevention focus and p=.870 for promotion focus). The negative association between women and advising and, to a lesser extent, monitoring, poses the question on whether a negative association also exists between women and engagement in monitoring and advising when these variables are peer rated rather than CEO rated. Interestingly, the results are even clearer for the peer-rated measures, which showed that females are significantly associated with advising (β=−.150; p=.000) and monitoring (β=−.330;
p=.001).
--- INSERT TABLE 8 ABOUT HERE
---
analysis instead of a polynomial one was used to assess the influence of fit. Hypothesis 5 predicts that prevention focus congruence between directors and CEOs has a positive effect on monitoring engagement. Although the hypothesis was originally rejected, the supplementary analysis shows that CEOs’ prevention focus moderates the relationship between directors’ high levels of prevention focus and monitoring engagement so that the effect becomes positive and significant (β=.190; p=.042). These results provide some support for Hypothesis 5. Concerning Hypothesis 6, which predicts a promotion congruence effect on advising, no significant effects could be found at either high or low levels of promotion focus.
--- INSERT TABLE 9ABOUT HERE
---
Finally, as directors do not generally operate absent of other directors, scholars have begun regarding regulatory focus theory as a multilevel phenomenon that can also occur at the board level. This is generally referred to as CRF (Johnson et al., 2015). Basing on this reasoning, the supplementary analysis explores how the CRF of the board influences board monitoring and board advising. To conduct the analysis, I aggregated the data to the organization level, resulting in a sample of 60 organizations. Both board monitoring and board advising are mean aggregated variables that reflect the opinions of directors and senior management. The constructs were originally derived from four items (monitoring Cronbach’s alpha=.8030) and three items (advising Cronbach’s alpha=.8470). These items were adapted from the works of Carpenter and Westphal (2001) and McDonald and Westphal (2010). To access the collective regulatory foci (i.e., promotion and prevention foci) of each board, I used the average promotion and the average prevention focus scores of the boards’ directors.
in which the effect is significantly negative, the effect is insignificant (β=.077; p=.593) at the board level. Thus, Hypothesis 1 is also not supported at the board level. Hypothesis 2 predicts a negative relationship between a promotion focus and monitoring. This hypothesis was only confirmed for high levels of promotion focus at the director level and similarly at the board level. Although the results are negative, they are insignificant (β=−.265; p=.090), thus not providing support for Hypothesis 2 at the board level. However, given the limited sample size, one can expect this hypothesis to be confirmed when more data are used. Hypothesis 3 predicts a positive relationship between a promotion focus and advising. At the director level, this hypothesis was not confirmed, and the results did not differ at the board level (β=−.074;
p=.677). Hypothesis 4 argues that a negative relationship exists between a prevention focus
and advising. Although this hypothesis was confirmed at the director level, the results at the board level indicate a positive relationship between prevention-focused directors and advising (β=.561; p=.000).
--- INSERT TABLE 10 ---
Evidently, the two polynomial hypotheses were also tested. To test the polynomial hypotheses, I used the congruence between the CRF of the boards and the regulatory focus of CEOs. Hypothesis 5 predicts a positive congruence effect between a prevention focus and monitoring. Again, this was not confirmed at the director level. In line with the suggestions of Edwards (1994), whether the second-order polynomial terms (i.e., X2, XY, and Y2) significantly improve the variance explained in the monitoring variable was first analyzed. Adding such terms does not significantly improve the explained variance (see Table 11: ΔR2=.0432; p=.476). However, as the absolute change value shows a rather large improvement
was continued. The congruence slope coefficient (a1=.08; p=.586) is insignificant, so the
findings do not provide support for Hypothesis 5 at the board level. ---
INSERT TABLE 11 --- --- INSERT FIGURE 2a ABOUT HERE
---
The sixth and final hypothesis predicted a positive congruence effect between a promotion focus and advising. Again, the results were not confirmed at the director level (see Table 14). To repeat the analysis at the board level, I initially analyzed the change in R-square resulting from the second-order polynomial terms. This change is insignificant (see Table 14: ΔR2=.0852; p=.1696). To gain a complete understanding of the final hypothesis, I analyzed the
congruence line. The results on the congruence line at the board level are negative and significant (a1=−2.91; p=.0440), thus contradicting Hypothesis 6 at the board level and
suggesting that promotion focus congruence between directors and CEOs leads to worse advising.
--- INSERT TABLE 14 --- --- INSERT FIGURE 2d ABOUT HERE
---
effects for the regulatory congruence analyses. Specifically, it was expected that a prevention fit negatively influences advising, whereas a promotion fit is negatively related to monitoring. Concerning the prevention fit effect on advising (see Table 13), the change in R-square is rather small and insignificant (ΔR2=.016; p=.824). Nonetheless, both the congruence line (a1=.520;
p=.011) and the incongruence line (a3=.570; p=.011) are significant and positive.
--- INSERT TABLE 13 ABOUT HERE
---
To gain a better understanding of the effect of prevention focus congruence on advising, I examined the response surfaces in detail. As in before, the congruence line (X=Y) corresponds to the line from the far-left corner to the near-right corner. Contrastingly, the incongruence line (X=−Y) corresponds to the line ranging from the far-right corner to the near-left corner. In contrast to what was previously expected, the response surface of Figure 2c indicates that CEOs’ prevention focus hardly influences the outcome. Overall, the effect is primarily driven by directors’ regulatory focus so that the higher the prevention focus is, the higher the board monitoring.
--- INSERT FIGURE 2c ABOUT HERE
---
For the effect of promotion focus congruence on monitoring (see Table 12), one could expect to find a negative effect. The change in R-square, brought about by the second-order polynomial terms, is rather small and insignificant (ΔR2=.039; p=.528). A negative congruence effect would be expected. However, the results are insignificant, and a negative effect between promotion fit and monitoring (a1=−1,87; p=.132) can be observed, discrediting the argument
--- INSERT TABLE 12 ABOUT HERE
--- --- INSERT FIGURE 2b ABOUT HERE
---
Discussion
While existing research on corporate governance has long embraced the notion that CEOs’ psychological traits can influence firm-level outcomes, research has primarily analyzed CEOs’ psychological traits in isolation and through indirect means (e.g., narcissism accessed through signature analysis [see Ham, Seybert, & Wang, 2018]). Additionally, previous research has often not considered the personality traits and core motivations of other senior management team members and directors within a firm. The current study aims to address these shortcomings by studying how the personality/motivational traits (i.e., regulatory focus) of directors, both in isolation of and in conjunction with those of CEOs, affect engagement in monitoring and advising. To develop the arguments, I drew on prior psychology work— particularly regulatory focus theory and interpersonal regulatory fit theory—UE theory, and the literature on board effectiveness.
prevention-focused individuals adapt more strongly to social norms (Zhang, Higgins, & Chen, 2011). In the context of this study, adaptability to social norms could imply that in a group in which the dominant regulatory focus2 is a promotion focus or the CEO is more promotion focused, the director perceives his/her task to be more promotion focused.
Another potential reason for the counterintuitive results of Hypothesis 1 could come from the fact that prevention-focused individuals are frequently associated with lower performance ratings (Wallace, Little, Hill, & Ridge, 2010). In a similar vein, most of the CEOs in the sample are primarily promotion focused (about 94%), so most directors who have a dominant prevention focus (about 30%) automatically experience an interpersonal regulatory misfit with the CEOs who provide the performance ratings. This is crucial, as Hamstra et al. (2013) have found that a regulatory fit leads to more favorable performance evaluations. This study potentially highlights that a regulatory misfit could lead to unfavorable interpersonal ratings.
Another potential reason for the lower performance ratings can be derived from the fact that prevention-focused individuals are generally more risk averse (e.g., Förster, Higgins, & Bianco, 2003; Grant & Higgins, 2003; Higgins, 1997; Higgins, 2002; Molden & Higgins, 2004). More recently, Gino and Margolis, (2011) have discovered in a series of experiments that individuals’ regulatory focus influences their risk perception. Specifically, their findings highlight that individuals with a promotion focus tend to exhibit more risk-seeking behaviors, whereas individuals with a prevention focus tend to avoid risk (Gino & Margolis, 2011). A strong prevention focus potentially leads individuals to perceive intervening and monitoring as
2 The dominant regulatory focus was computed following established procedures. Specifically, the prevention
socially risky. Accordingly, prevention-focused individuals are potentially hesitant to intervene early, as their engagement in monitoring might be rated worse.
Additionally, one cannot neglect that methodological issues might have played a role in the counterintuitive results because the two dependent variables are highly correlated. Accordingly, the unique variance, which is present in each dependent variable, may not have been sufficient to find diverse effects on monitoring and advising.
As gender affects the results, it has also been assumed that a potential relationship exists between female directors and a prevention focus. This may have resulted in a negative association between a prevention focus and monitoring. Although considerable evidence shows that females are more prevention focused than males are (McKay-Nesbitt, Bhatnagar, & Smith, 2013), the study’s data do not reveal such an association between females and a prevention focus. Not all females are, by default, prevention focused, and accordingly, it is not unexpected that this relationship could not be observed. Prevention-focused females also often receive negative feedback on their preventive behavior, so a prevention focus might not be expressed (Ellemers & Rink, 2016). From this reasoning, it cannot be fully ruled out that the negative effects of gender contribute to the negative association between a prevention focus and monitoring.
One of the most surprising results of the study is that the effect of directors’ promotion focus on engagement in advising is highly insignificant. Yet, supplementary analysis has revealed that the effect is negative and significant at high levels of promotion focus (β=−.220;
p=.047), indicating that in contrast to what was hypothesized, the effect of promotion focus on
advising is negative at high levels of promotion focus. As these findings contradict the dominant logic, future research should investigate the observed phenomenon. Highly promotion-focused directors are likely to be eager to give the potentially best advice, so one possible reason for these findings could be that their eagerness hampers their overall engagement in advising.
Unsurprisingly, Hypothesis 4, which predicts that a prevention focus is associated with worse performance in providing advice to CEOs, is confirmed (β=−.120; p=.040). To better understand the findings, I additionally investigated how the effects differ for high and low prevention foci. In line with the hypothesis, I found a significant and negative effect at high levels of prevention focus (β=−.230; p=.038), whereas the effect is positive and significant at low levels of prevention focus (β=.230; p=.038). This result clearly suggests that low levels of prevention focus are beneficial for advising, whereas high levels are detrimental to engagement in advising.
2011) and implies some level of distrust between parties. Accordingly, the positive relationship created through a regulatory fit can serve as an obstacle to the performance of monitoring tasks. Another reason why an interpersonal prevention fit may actually have a negative effect on monitoring is that either the director sees little need to monitor the CEO closely, or the CEO feels capable of self-monitoring and therefore perceives the director’s monitoring as unnecessary. Yet, the supplementary analysis has shown that CEOs’ prevention focus moderates the relationship between a prevention focus and monitoring at high levels of directors’ prevention focus so that it becomes positive. Taken together, these findings indicate that although no congruence effect exists, the relationship between directors’ high prevention focus and monitoring is moderated by CEOs’ prevention focus.
Hypothesis 6 predicts a positive effect of promotion congruence on advising. Yet, the main analysis and the supplementary analysis have shown that the effects are insignificant. One possible reason for this might be that CEOs with a strong promotion focus need less advice because they have a natural tendency to explore every opportunity, which implies that directors may not have many options to provide new advice to them.
Given the unexpected results, the research also accessed the hypotheses at the board level. Although prior research has primarily focused on a regulatory focus at the individual level (Johnson et al., 2015), group research has shown that people in a group tend to adapt to and conform with the norms of the groups they operate in. Accordingly, groups can develop a shared understanding of tasks and can even influence what is discussed and which decisions are taken (Florack & Hartmann, 2007). On the basis of this reasoning, exploring if the counterintuitive or insignificant results of the study are better investigated at the board level rather than at the director level is crucial.
speaking, the findings indicate that the board level’s direct effects of a regulatory focus are less pronounced than those at the director level. One explanation for these results is the limited sample size. Furthermore, exploring the reason for why a prevention focus leads to improvements in advising remains essential. The positive effect of board prevention focus on advising potentially comes from industry effects, as most organizations are non-profit. Consequently, prevention-focused advising might be highly valued by CEOs and, therefore, more frequently given. Nonetheless, future research should investigate why this positive relationship is only present at the board level but not at the director level.
The congruence effects at the board level are insignificant for monitoring, whereas both prevention and promotion congruence have a significant influence on advising. The insignificance of congruence effects can be potentially explained by the fact that directors are legally liable for their action and, in particular, for how well they protect shareholders, as monitoring is the only governance mechanism that allows directors to direct and adjust senior management’s behaviors. Given the personal and professional repercussions that bad monitoring can have for directors, directors are keen to fulfill this task notwithstanding their regulatory focus. Although advising is also a crucial part of governance, the task usually comes with fewer threats. Accordingly, more variance in the responses for advising can be expected. In line with this reasoning, the paper found a positive effect of prevention focus congruence on advising and a negative effect of promotion focus congruence.
in line with the suggestion made by Gamache et al. (2015) that directors should take into account CEOs’ regulatory focus when directing them. This comes from the consideration that prevention-focused directors probably need more advice and encouragement to take on riskier ventures, whereas promotion-focused CEOs may need more vigilant oversight and monitoring by directors. The positive congruence effect of a prevention focus on advising could be driven by the sample selection, which primarily includes non-profit organizations. In these types of organizations, the advice given by highly prevention-focused individuals is likely to be more appreciated because of the greater financial constraints and public pressure often present in non-profit organizations. Therefore, future research needs to expand these findings to other industries in order to create generalizable results.
The discussion above highlights that the results of the hypotheses are quite different at the individual and board levels. One potential reason for this is that, as mentioned previously, monitoring and advising are highly correlated (see Table 1) at the individual level and share more than half of their variance. At the board level, the correlation between monitoring and advising is much lower (r=.541), which implies that the two constructs only share about 30% of the variance, whereas the shared variance at the individual level is just above 50%. This fact indicates that enough distinct variance exists at the board level to find diverse effects on monitoring and advising.
the results indicate the importance of directors’ prevention focus motivation to understand governance effectiveness (i.e., monitoring and advising).
Another contribution of this study concerns interpersonal regulatory fit. Although the findings indicate that the effects of prevention and promotion congruence on either monitoring or advising are insignificant at the director–CEO level, the supplementary analysis has revealed that both prevention and promotion congruence significantly affect board advising at the board–CEO level. This result provides some indication that governance effectiveness is primarily a board-level phenomenon and that a single individual struggles to improve governance effectiveness.
In general, this research contributes to the development of micro-level antecedents to governance effectiveness. This is crucial for the literature, as the firm level cannot be fully understood without considering the individuals who make up organizations (Felin & Foss, 2005). Ultimately, the findings help enrich the understanding of when and why governance failures occur.
Finally, this study carries implications for practice. First, directors, who are aware of their regulatory focus, become better able at grasping their natural tendency to be extremely vigilant or more opportunity oriented, so they can learn to better balance their traits depending on the person or situation involved. This awareness can help prevention-focused individuals avoid negative performance ratings, as they can decrease their natural tendencies to levels that are appropriate in a given situation. This information might be especially helpful for female directors, as it may assist them in avoiding double scrutiny in performance ratings.
Limitations and future research
As in any study, several limitations need to be acknowledged. First, caution is warranted when generalizing the results because of the study design. The reason for this problem is fourfold. To begin with (a), the data used in this study are cross-sectional, which formally implies that causal relationships cannot be inferred, yet the model implies a casual order. To allow stronger inferences about the casual order of the proposed model, future research should further analyze this relationship by using a longitudinal study or an experimental design. Nonetheless, it seems rather unlikely for adverse effects to occur in the study because a regulatory focus is a relatively stable trait (Higgins, 1997; Higgins et al., 2001; Strauman, 1996) that cannot be influenced by monitoring or advising in the board. Additionally, the combination of individual, interpersonal, and peer-rated measures makes it even more unlikely that reversed causality exists. Given the difficulty to gain access to longitudinal or experimental data from boards and the strong conceptual reasons for the implied causality, the benefits of either a longitudinal study or experiments are limited because of the costs involved.
Another limitation (b) that affects the generalization of the findings is that all firms and individuals who participated in the study are from the Netherlands. As stated previously, Dutch firms are required by law to have a two-tier board system in place (Hooghiemstra & Van Manen, 2004). Although the tasks of supervisory directors in a two-tier system and those of non-executive directors in one-tier systems are very similar, the results can possibly differ in countries that rely on a one-tier board system.
The final limitation (d) that contributes to the limited generalizability of the findings is concerned with the supplementary analysis at the board level. Specifically, only 60 observations were available at the board level. Usually, scholars studying the functioning of boards use large samples and proxies. Contrastingly, this research uses micro-level data at the board level, and it does not use proxies. Future research could aim to use proxy measures and a larger sample.
Second, some of the counterintuitive results could not be empirically examined within the present study. This included, among others, the negative influence of a high prevention focus on monitoring, as well as the negative congruence effect of a prevention focus on monitoring. A prevention focus has been theorized, in general, to lead to lower interpersonal performance ratings. This suspicion is somewhat strengthened by the fact that one can observe a positive congruence effect between a prevention fit and monitoring at the board level. Future research could aim to find experimental evidence for this phenomenon.
Third, the supplementary board-level analysis assumed that each directors’ regulatory focus has the same influence on the CRF of the board. However, in boards and in any other groups, formal (e.g., chair, vice chair, or committee chair) and informal hierarchies (e.g., expertise and status) emerge (Anderson & Brown, 2010). Additionally, leaders can influence the work regulatory focus of followers (Neubert et al., 2008). In the context of boards, those directors further up in the hierarchy can influence others’ work regulatory focus and therewith disproportionally influence the CRF of the board. Future research would benefit from carefully analyzing how the CRF in boards develops and the extent to which directors actually influence one another, given the limited time period they actually work together. Basing on this, one could build a more reliable measure of collective regulatory focus.
