As owners continue to cede decision-making powers to executives in response to increasingly complex and dynamic business realities, the board has been cast in the role of guardian of shareholder welfare (Fama and Jensen, 1983). Board monitoring may help resolve the agency problem between owners and top executives, but considerable variance exists in the extent to which boards actually influence strategic decision making (Johnson, Daily and Ellstrand, 1996; Carpenter and Westphal, 2001). The effectiveness of board monitoring is mainly limited by information asymmetries and goal incongruence in the board–management relationship (Jensen and Meckling, 1976). The situations become more critical when management decisions pertain to uncertain or risky investments, such as new product development which is by nature new to either the organization or the market (Wade, O’Reilly and Chandratat, 1990; Damanpour, 1991). While firm innovativeness is considered in the realm of corporate governance, a firm's pursuit of innovation particularly requires aligning the interests of board directors and CEO, even though these two parties, which command resource allocations and ultimately are responsible for firm success, often conflict.
Although prescriptive advice about the need for harmonious working relationships abounds (Stewart, 1991; Roberts and Stiles, 1999), empirical inquiry into the extent to which board–CEO relationships shape strategic outcomes remains relatively lacking (McNulty and Pettigrew, 1999). In a conceptual framework of board building, Nadler (2004) argues that boards vary along a continuum, from least to most engaged, and that corporate scandals often expose boards as either too passive or too involved. Although little conclusive evidence indicates that structural power increases board involvement in strategic decision making (Westphal, 1998), a power bias for either the board or management risks potentially dysfunctional rivalries and fragmentation (Hambrick and Finkelstein, 1995). On the other hand, agency conflict may take place differently in the presence of social ties between boards and top executives (Daily, Dalton and Rajagopalan, 2003). Though empirical research offers no comprehensive consideration thus far, social ties may change a board's willingness or ability to exercise monitoring or provide expertise, and thus shape the effects of interest alignment on new product development decisions (Lynall, Golden and Hillman, 2003; Maitlis, 2005). Accordingly, from the perspectives of sociopolitical and sociopsychological forces rooted in the board–CEO relationships in Taiwanese firms, this study departs from the traditional agency view holding that a board's lack of independence from management weakens its contribution to strategic decision making (Kang, Cheng and Gray, 2007). Instead, this research posits that fine-tuned board–CEO relationships, in terms of power balance and social ties, will serve as an essential enabling context for managerial risk taking, and thus enhance firm innovativeness.
Given governance does not function in isolation and must adapt to changing conditions (Strebel, 2004), some institutional characteristics distinguish Asian corporate governance from that in the West. For example, in Taiwan, as in other East Asian countries with civil law jurisdictions, company board structures adopt a two-tiered design (i.e., distinct managing and supervisory functions), which generally has not been considered an effective control mechanism because of the varying degrees of board involvement and independence (Solomon, Lin, Norton and Solomon, 2003). In this sense, the contingent efficacy of any governance arrangement should reflect the context in which it functions. According to strategic management literature, environmental characteristics represent all external forces on firm actions (Bourgeois, 1980; Dess and Beard, 1984; Sharfman and Dean, 1991), and interorganizational characteristics focus on external links, assuming that firms are embedded within networks of relationships that provide opportunities for and constraints on firm behavior (Westphal, 1998; Leenders and Gabbay, 1999). Both notions not only yield direct implications for firm performance, but also play moderating roles in a number of organizational relationships (Prescott, 1986; Zaheer and Bell, 2005). Furthermore, in response to criticism of firm-level-only approaches that overlook contextual factors that might influence and constrain firm behavior, this study proposes a contingency view to answer the recent calls for micro–macro links in strategy research (e.g., Hackman, 2003); that is, it is suggested that the innovation-determining effects of board–CEO relationships depend on environmental instability and board interlocks. Therefore, in addition to addressing the critical role of the board/CEO relationships in shaping firm innovativeness, the study aims to highlight the context-dependent nature of governance arrangements.
This study thereby contributes to corporate governance knowledge by bridging the gap between board–CEO relationships and strategic outcomes in terms of new product introduction. First, the findings clarify the implications of rising tensions between proponents of intervening approaches and those that propose passive approaches to board governance. Debates persist about whether detached outsiders or knowledgeable insiders should dominate boards, and whether social dependence between the board and executives compromises the board's strategy-making effectiveness. This study argues for a fundamental reframing of governance issues and moves beyond either/or thinking. Polarizing the governance prescription to stress the dominance of one party rather than the complementarity of both raises defenses, impedes mutual learning, and hinders managerial risk readiness. Second, prior studies of corporate governance overlook the context-bounded nature of their fineings, leading to the conclusion that some governance régimes are universally effective. Yet, researchers increasingly call for understanding that moves beyond such context-free logic. By showing that the efficacy of board–CEO relationships depends on environmental and interorganizational characteristics, this study explores when specific governance arrangements promote firm innovativeness. Third, the performance implications of corporate governance have been widely documented – primarily in industrialized economies. The corporate landscape in Taiwan is an archetypical example of the newly industrializing economies in East Asia, so the use of Taiwanese data enables a better assessment of the generalizability of prior conclusions about the effect of aligned interests, and explains the governance–innovation relationship within the context of emerging economies, as has not been considered in previous research.
The paper proceeds as follows: I begin by declaring the research goal and delineating its significance. I then bring theoretical relevance to the research subject and develop testable hypotheses. The third section reports on the sample and measures. The subsequent section presents the empirical results, and the final section concludes, identifies some limitations, and develops managerial implications.
THEORIES AND HYPOTHESES
Because the board resides at the apex of a firm, its directors are expected to provide critical judgment about management performance, which requires in-depth knowledge of and intimacy with the affairs of the corporation (Keenan and Aggestam, 2001). Directors also must provide independent judgment, which requires their detachment and distance from the top management team (Demb and Neubauer, 1992). Although the nature of the relationship between the board and top executives, as well as its influences on organizational outcomes, remains inconclusive, several theoretical strands shed some light on the subject. Among them, agency theory represents perhaps the most recognized. According to this perspective, boards of directors exist to monitor managers on behalf of shareholders and ensure that their interests are pursued (Jensen and Meckling, 1976), because shareholders and executives exhibit fundamentally different relationships with organizational processes and outcomes (Daily et al., 2003). While ordinary shareholders may be considered risk neutral, because they can diversify their risks through their portfolios (Wiseman and Gomez-Mejia, 1998), executives cannot diversify their risks as easily. In addition, innovative activities often promise uncertain outcomes, and innovation failures can not only depress a firm's short-term performance, and thus lower management compensation, but also damage executives’ reputations and increase their risk of unemployment. Thus, executives likely are risk averse. To counter this aversion, shareholders may resort to corporate governance, using monitoring or incentives, to align their risk differentials. Although board monitoring can help resolve the agency conflict, at least partially, its effectiveness also is limited by potential information asymmetries. Rather than pointing directly to the board–CEO relationship, agency theory focuses more on the importance of board power as a means to cure the agency problem, but it leaves the dangers of overcentralized power with the board.
In contrast to the focus of agency theory – that is, the board as a shareholder guardian and monitor of managerial actions – the board also can be viewed as a conduit for managing resource dependencies within social networks. By considering firms open and dependant systems that rely on the environment for support, resource dependence theory (Pfeffer and Salancik, 1978; Granovetter, 1985) argues that an organization takes either a buffering or a bridging approach in response to resource scarcity in the external environment. According to this point of view, directors of boards can co-opt important external organizations and provide different linkages and resources to the firm. Viewing the board as a provider of resources rather than an evaluator of management, the resource dependence logic also suggests that a board's function relates directly to firm performance. A resource, or anything that could be thought of as a strength or weakness of a given firm (Wernerfelt, 1984), helps reduce dependency of the focal firm on external contigencies, diminish its uncertainty, lower transaction costs, and ultimately aid in firm survival.
Regarding the roles associated with the resource provision function of the board, Johnson et al. (1996) define a “service role,” such as when directors advise CEOs or other executives about managerial issues and actively initiate or formulate strategy, versus a “resource dependence role,” which occurs when directors facilitate the acquisition of resources critical to the firm's success by serving a legitimizing function. Both roles represent part of the same general provision of resources logic (Pfeffer and Salancik, 1978). In the context of a social network, this view not only echoes the importance of the interorganizational environment for reputation, trust, reciprocity, and mutual interdependence (Larson, 1992), but also considers the predictable paths (based on interlocking directorates) that might acquire these resources or facilitate the flow of resources within the network.
In addition, other theoretical perspectives shed some light on the roles and interaction of the board and CEO within a company. For example, with an emphasis on the socially constructed nature of role playing, role theory argues that a complementary board–CEO relationship involves the fit between abilities and responsibilities, the environmental and organizational contexts, and the personal relationships and expectations that surround these top positions (Roberts, 2002). Psychoanalytic literature on leadership (e.g., Kets de Vries, 1991; Garnsey and Roberts, 1996) also explores the extent to which the two parties’ interactions send signals to the wider organization. With regard to the influences of prevailing institutionalized norms, institutional theory (Meyer and Rowan, 1977; DiMaggio and Powell, 1983) holds that the persistence of certain governance arrangements, such as the division of labor between the board and CEO, represents a way to secure legitimacy rather than a process driven by rational considerations.
As the debates persist about whether the board should monitor or empower management, tensions between the control and collaborative approaches to governance expand (Dalton, Daily, Ellstrand and Johnson, 1998; Sundaramurthy, 2000). Understanding the nature of the board–CEO relationship is not merely a matter of delineating formal job specifications or defining roles, which would fail to capture the nuances of complementarity. Instead, the nature of the board–CEO relationship and its effects on strategic decision making entails a complex, multidimensional organizational phenomenon that may not be comprehensible within a single theoretical perspective (Finkelstein and Hambrick, 1996). Considering a number of organizational paradoxes, such as the coexistence of authority and democracy, efficiency and creativity, and discipline and autonomy (Lewis, 2000), Sundaramurthy and Lewis (2003) suggest that researchers should accommodate different theories to elaborate on the underlying tensions in the corporate governance and emphasize the merits of complementarity. The aforementioned theories may have been applied, respectively, to explaining certain governance phenomena, but questions remain as to their varying explanation and boundary conditions. Each theory may sensitize us to one unique aspect of the board–CEO relationship, but a multitheoretical view would be more appropriate in a strategic risk-taking context. In this vein, the sociopolitical aspect of structural power balance between the board and CEO is just one part of the story; the sociopsychological aspect, or a concern with the social dependence between the parties, represents another.
This study assumes that divergence in the risk profiles of shareholders and managers affects only the qualitative aspects of R&D allocation decisions, such that management prefers to steer resources toward more near-term, less risky, more incremental, less profound R&D projects that offer smaller but more certain results. In this sense, management may pursue innovation, but prefer a more “exploitative” R&D strategy, whereas shareholders prefer an “explorative” R&D strategy (Hill and Snell, 1989). Therefore, innovation outcomes in terms of new product introduction, which by its very nature is disruptive because of its novelty to the organization and the market, require more managerial risk taking. Such an effort, in turn, depends on the extent to which managers’ risk preferences align with those of shareholders through governance arrangements.
The Formal Board–CEO Relationship (Board Involvement in Company Decision Making)
Although boards of directors may play multiple roles in organizations, prior literature concerns mainly with the explicit dimensions of a board, such as its size, background, equity ownership, composition, and independence. As governance remedies based on such “usual suspects” generally do not work as expected (Finkelstein and Mooney, 2003), the focus has shifted to the governance process, which suggests that the nature of the interactions between the board and executives may influence board effectiveness with regard to fulfilling the key roles of monitoring, advising, and counseling top executives (Forbes and Milliken, 1999).
Governance literature shows considerable variance in the degree to which boards have actual impacts on strategic decision making (Wade et al., 1990), but a critical issue involves the level of engagement of the board in influencing management decisions and firm strategy (Montgomery and Kaufman, 2003). Pettigrew and McNulty (1995) draw a broad distinction between what they call “minimalist” boards, which minimize the impact of part-time board members, and “maximalist” boards, in which nonexecutives shape the content, context, and conduct of the strategy along with executives. However, this dichotomization does not lead to board effectiveness, as they argue. Nadler (2004) also proposes a conceptual framework of board building, arguing that boards vary along a continuum from least involved (i.e., passive or certifying board, minimal activities, participation at managers’ discretion) to most involved (i.e., intervening or operating board, makes all key decisions for management). Yet, despite the increasing importance of the strategic involvement of boards, the levels of board involvement may change according to the relevant issues and circumstances.
Board involvement enhances governance through vigilant controls on management (Fiegener, 2005). Empirical research points to the lack of board power to explain limited board effectiveness (Mallette and Fowler, 1992; Boeker and Goodstein, 1993; Baliga, Moyer and Rao, 1996). For example, Kimberly and Zajac (1988) find that a highly involved board fosters greater confidence in managerial decision making, and relaxes tensions between directors and CEOs. Johnson, Hoskisson and Hitt (1993) show that powerful boards (i.e., outside directors hold significant ownership and are heavily represented) are more involved in corporate restructuring decisions; however, the presence of powerful executives (i.e., high levels of executive ownership, lengthy tenure) tends to forestall board involvement in decisions. Lorsch (1996) also espouses that the modern board has progressed from defending the corporate status quo to becoming a force for change. For instance, some boards position themselves as providers of resources and expertise ex ante rather than as performance evaluators of management ex post. Such strategic control ensures that executives proactively examine appropriate alternatives, and creatively respond to opportunities when they arise (Shimizu and Hitt, 2004).
In contrast, strong control by an intervening board may cause unexpected outcomes, as the expropriation hypothesis proposes. Shleifer and Vishny (1997) argue that agency problems may stem from the conflict between controlling owners and minority shareholders, instead of that between managers and dispersed shareholders. Therefore, powerful boards are costly because majority owners can redistribute wealth – in both efficient and inefficient ways – from other minority shareholders, whose interests need not coincide. Denis, Denis and Sarin (1997) also suggest that blockholders, who are usually powerful board members, seek not only to increase firm value (shared benefits of control), but also enjoy benefits unavailable to other shareholders (private benefits of control). Such private benefits may come at the expense of other, smaller shareholders.
Although little conclusive evidence indicates that structural power increases board involvement in strategic decision making (Westphal, 1998), a power bias toward either the board or management risks potentially dysfunctional rivalries or fragmentation. According to Nadler's (2004) observations, corporate scandals expose boards as either too passive or too involved and indulgent. Given that varying levels of board involvement present a dilemma between oversight and micro-management, Finkelstein and Mooney (2003) suggest that boards should maintain an “appropriate level of involvement” in company decision making, to avoid encroaching on management's role. Because a governing body should operate in an atmosphere of constructive discontent, and tap the positive tension that stems from shared values but distinct accountabilities (Sundaramurthy and Lewis, 2003), these competing arguments, pertaining to the convergence-of-interest effect and entrenchment effect, suggest a nonlinear relationship between board control over top executives and performance outcomes. As the influence of board involvement in company decision making on the risk-alignment effect may follow similar theoretical logic, this study argues that new product introduction, as a result of managerial risk taking, first increases and then decreases as the structural power of the board increases.
Hypothesis 1: Board involvement in company decision making has an inverted U-shaped relationship to a firm's performance of new product introduction.
The Moderating Effect of Market Instability
Environmental factors outside the control of managers usually influence the success of their strategic decisions. Organizational environments may be characterized in terms of their components, which include customers, competitors, and suppliers, and their attributes, such as instability, munificence, and complexity (Bourgeois, 1980). Among these environmental characteristics, instability entails the extent to which market demand and underlying technology change rapidly in a given industry (Dess and Beard, 1984; Sharfman and Dean, 1991); it thus represents a critical antecedent of organizational structure, strategy, and outcomes (Keats and Hitt, 1988). Often synonymous with uncertainty, dynamism, or volatility, environmental instability alters the effectiveness of managerial decision making (Li and Simerly, 1998), which is viewed as a moderator for the innovation-determining effects of board–CEO relationships in this study. For all parties involved in governance, an increase in environmental instability not only results in a greater inability to assess both the present and future state of the environment accurately, but also highlights the indispensability of managerial discretion, in which top executives are required to exercise latitude in their actions based on their professionalism and experiences in handling such circumstances (Hambrick and Abrahamson, 1995).
Prior literature often considers the need for decision-making processes to deal effectively with environmental constraints. For example, upper echelons literature suggests that the characteristics of top executive teams are important to the extent to which they are aware of or respond to changes in environments (Hambrick and Mason, 1984; Sutcliffe, 1994). Literature on environmental scanning (Daft, Sormunen and Parks, 1988) suggests that structural characteristics influence information acquisition processes. In addition, to understand whether effective managerial action is better served by rational analysis or creative intuition, Sadler-Smith (2004) proposes the moderating role of environmental instability on the performance effect of managers’ intuitive decisions, and concludes higher environmental uncertainty as a suitable context for managerial freedom of choices.
Although the need to align managerial preferences with the interests of owners is of paramount importance, environmental unpredictability augments the variance of decision makers’ perceived uncertainty, and generates disputes about goals and means, making it difficult to centralize control over goal formation and attainment. Similarly, environmental instability is incompatible with autocratic (i.e., centralized) governance, because the rigidities of board domination, particularly by outside directors, conflict with rapid adaptation (Knoke, 1990). Dean and Shareman (1996) argue that environmental instability positively conditions the relationship between managerial power, in terms of information analysis and filtering, and decision-making effectiveness: the relationship is stronger in unstable environments than in stable ones. Jarley, Fiorito and Delaney (1997) also argue that environmental instability relates negatively to the centralized control of the board over goal formation.
Because managerial discretion fundamentally represents a form of strategic response to environments, top executives who fail to adapt to, or lack the latitude to respond to, environmental changes can achieve only nonviable outcomes in unstable environments. In contrast, executives in stable situations already have an experience-based understanding of their environment, and thus less need to engage in risk-taking behavior. As high levels of market instability can be seen as a context for managerial discretion, necessitating CEOs’ freedom of choosing a wide array of potential courses of action in major decision domains. The argument about the influence of market instability now must be integrated with the “board involvement–firm innovativeness” relationship. Thus, the level of market instability is hypothesized to negatively moderate the curvilinear relationship delineated in Hypothesis 1.
Hypothesis 2: Among firms subject to greater market instability, board involvement in company decision making is associated with a poorer performance of new product introduction.
The Informal Board–CEO Relationship (Social Ties between the Board and CEO)
The structural power balance between the board and CEO recognizes the divergent risk preferences of these two major parties, which can be aligned by sociopolitical forces. However, this perspective is contrasted with sociopsychological ties. Although boards contribute to strategy by exercising oversight and control, they also provide expert advice and counsel in their administrative roles (Pfeffer and Salancik, 1978). This function refers to the ability of the board to integrate resources into the firm (Hillman and Dalziel, 2003), assuming a definition of resources as “anything that could be thought of as a strength or weakness of a given firm” (Wernerfelt, 1984). However, empirical research neither explicitly models resource-providing relations nor examines how social ties, such as friendship and family relationships, may influence a board's ability to undertake this function (Lynall et al., 2003; Maitlis, 2005).
Agency effects may function differently in the context of family firms (Daily et al., 2003), so prior findings pertaining to board effectiveness in non-family businesses may not generalize to settings that feature strong social ties between the board and CEO. According to literature on advice seeking, an advice seeker's status impairs the effectiveness of this seeking. In other words, people believe that others will view their need for assistance as a confession of uncertainty or dependence, implying that they are not fully competent or self-reliant (Rosen, 1983). Moreover, seeking advice from a superior may entail disclosing problems and admitting personal limitations, which in turn could force the advice seeker to relinquish power. This concern about losing status and disclosing information inhibits executives from seeking advice from the board. However, personal relationships encourage them to seek advice by creating a sense of social security that reduces perceived embarrassment and risk (Anderson and Williams, 1996). Dalton et al. (1998) point out that directors connected to the firm through family relationships or as stakeholders often experience a greater incentive to provide advice and counsel, call on connections with other organizations, encourage communication flows, and act to improve the external image of the firm. In a study of the board's function in advice provision, Westphal (1999) concludes that social ties between board members and executives foster a more collaborative decision-making environment and promote the board's strategic involvement. Thus, according to his collaborative board model, social ties help increase the prominence of advisory interactions as a form of involvement and, particularly through links with outside directors, increase the propensity of top executives to solicit advice on strategic issues.
In contrast to these positive effects, boards that lack social ties to management are less likely to exert control over strategic decision making on behalf of their shareholders (Boeker and Goodstein, 1993; Baliga et al., 1996). For example, boards with more outside directors are less likely than those with more insiders to collude with managers to expropriate residual claimants, so the traditional agency perspective always highlights the importance of board independence (Fama and Jensen, 1983). In a similar vein, close ties between the board and CEO may weaken board effectiveness by reducing objective and independent monitoring activities. Claiming that a lack of social independence can compromise board monitoring in the strategy-making process, Wade et al. (1990) argue that many CEOs intend to keep their boards passive and uninvolved in strategic decision making, or else pack the boards with their supporters. Board directors also are less likely to engage in vigilant monitoring and exert control over top executives with whom they have close personal connections (e.g., Fredrickson, Hambrick and Baumrin, 1988; Walsh and Seward, 1990). Considering behavioral processes and dynamics in board–management relationships, Westphal (1999) hypothesizes that the friendship between the board and CEO is negatively associated with board effectiveness. Strong social connections can likewise bias the board's perceptions of CEO, which hampers its ability to monitor and discipline incompetent CEOs. Schulze, Lubatkin, Dino and Buchholtz (2001) also argue that family ownership does not necessarily minimize the agency cost of fractional ownership but, in some cases, can actually exacerbate it. Finally, Gomez-Mejia, Larraza-Kintana and Makri (2003) contend that family ties appear to shield CEOs from uncontrollable (systematic) business risk, leading to greater organizational complacency.
Although social connections to the top executive team may create an incentive for the board to provide resources, such intimate ties also may represent a disincentive for vigilant and unbiased monitoring (Dalton et al., 1998). Even if board–CEO ties temper self-interest and engender loyalty, commitment, and a long-term perspective, altruism of this kind can alter the incentive structure of a firm, such that some agency benefits get offset by free riding and other agency problems (Schulze, Lubatkin and Dino, 2003). In light of these countervailing forces, this study views these perspectives as not necessarily incongruent, but rather as operating at different levels of board–CEO ties. Social ties, on the one hand, bring about mutual trust, facilitate collaboration, complement rational controls, and serve as important lubricants for advice/counsel provision between the board and CEO. On the other hand, overly close personal connections may lead to extreme cohesion and enhance group thinking. Because managerial risk taking is a function of board governance, which itself is influenced by social ties between the two parties, this study predicts a curvilinear relationship between the board–CEO ties and the success of new product introduction. Specifically, an intermediate level of social ties in any given governance setting may be optimal for firm innovativeness.
Hypothesis 3: The level of social ties, in terms of family and friendship ties, between the board and CEO has an inverted U-shaped relationship to a firm's performance of new product introduction.
The Moderating Effect of Interlocking Directorates
The persistent challenges faced by board directors who want to make meaningful contributions to strategy involve not only their structural power (Demb and Neubauer, 1992; Westphal and Zajac, 1997; Westphal, 1999), but also whether they have suitable knowledge or information to contribute. Pfeffer and Salancik's (1978) influential discussion of the different possible functions performed by boards distinguishes two roles: an administrative body and a link between the organization and its environment. Although serving on the boards of multiple companies makes it difficult to gain an adequate understanding of the issues facing any one firm, such sharing of board directors can provide access to valuable information (Judge and Zeithaml, 1992; Daily and Dalton, 1994). Directors involved in corporate interlocks, or networks of interlocking directorates, not only diffuse organizational practices (Ornstein, 1980), but also create dynamics between the board and CEO because of a sense of unity among members of this elite class (Mizruchi and Stearns, 1988).
Directors’ appointments to other boards may dilute their capacity to contribute to decision making at a focal board, but existing literature suggests that directors involved in other boards provide an important source of information about business practices and policies (cf. Palmer, Jennings, and Zhou, 1993; Mizruchi, 1996). For example, Useem (1982) shows executives use cross-board appointments to scan the environment for timely and pertinent information, and concludes that direct involvement in other companies’ affairs provides a platform for learning and offers more strategic alternatives. Similarly, directors in the interlocking boards are better able to use such opportunities to learn firsthand about the efficacy of different practices, and how to implement them properly by observing the consequences of other managers’ decisions (Haunschild, 1993). Such learning is particularly helpful in their monitoring role and advising management.
Thus, according to the sociocognitive perspective, which considers variation in directors’ networks of appointments and the performance implication of such network ties, directorate interlocks provide more strategic knowledge and resources to monitoring and advising managers, and thus positively moderate the curvilinear relationship posited in Hypothesis 3.
Hypothesis 4: Among firms with higher levels of board interlocks, the social ties of boards–CEOs are associated with a better performance of new product introduction.
DATA AND MEASURES
To test the hypotheses, a survey of board–CEO relationships and new product introduction in Taiwan provides pertinent data. This research approach includes a broad sample of firms and industries in an attempt to maximize variation of all variables and increase the generalizability of the findings. The sampling frame, constructed from the databases of the Taiwan Credit Information Service Incorporation (TCISI), offers relatively complete coverage of Taiwanese industrial firms. A random stratified sampling method selects 1,000 firms from the TCISI database.
The self-reported survey was administered using a modification of Barringer and Bluedorn's (1999) procedure. After completion of the pilot study, a revised survey instrument was prepared and mailed to the CEO or a member of the top management team who was familiar with the board characteristics and board process of each sample firm. Two weeks later, a second copy of the survey was sent to the nonrespondents, along with follow-up letters. Responses with missing data or unclear or contradictory answers that could not be reconciled through follow-up telephone calls were removed from the sample. This approach leaves a total of 198 valid questionnaire replies, for a response rate of 19.8 per cent. The firms that responded to the survey represented a broad cross-section of Taiwanese, publicly held companies, ranging in size from 34 to 16,595 employees and operating across industries, including both traditional and high-tech sectors. The mean number of employees of the responding firms was 754.
To minimize the threat of self-report biases, two tests serve to check the validity of survey data. First, Harman's one-factor test reveals the presence of any common method variance by including the variables in a factor analysis (Podsakoff and Organ, 1986). The test yields 23 factors with eigenvalues greater than 1, and no single factor is dominant; therefore, common method variance is not a significant problem. Secondly, to assess the presence of nonresponse bias in the data, firms that responded to the survey were compared with those that did not on the basis of three characteristics: firm sales, number of employees, and 2005 returns on sales. According to the Kolmogorov–Smirnov test (Siegel and Castellan, 1988), no significant difference appears in the distribution of respondents and nonrespondents for any given variable. The statistically significant results (p = .42–.68) provide consistent evidence across multiple variables that respondents and nonrespondents come from the same population.
New Product Introduction
Most innovations cannot create economic value until they have been put to use and introduced into the market. Therefore, new product introduction serves as an important measure of firm innovativeness, in that it indicates the potential commercial significance of the firm's innovation activities (Katila, 2002). In a meta-analysis of innovation studies, Damanpour (1991) finds that this count provides a robust measure of innovation over a wide range of research settings. Because archival sources of performance data at the new product level are unavailable, this study relies on the surveyed firms to provide objective figures of new product introduction (Autio, Sapienza and Almeida, 2000; Li and Atuahene-Gima, 2002). Specifically, respondents report the percentage of total annual sales derived from products or services introduced during the past three years (i.e., 2003–2005). New products or services are defined as those with either unprecedented performance features or familiar features that offer significant improvements in either performance or cost and that, thus, transform existing markets or create new ones. The average figure over three years for each firm then serves as the measure of innovation performance; it correlates significantly with other commonly observed proxies of firm innovativeness, such as average percentage of sales spent on R&D (r = 0.45, p < .01), number of personnel assigned to R&D (r = 0.56, p < .01), and sales growth (r = 0.43, p < .01).
Board involvement refers to the extent to which the board's decisions dominate a firm's decision making. Based on prior literature (Pettigrew and McNulty, 1995; Nadler, 2004), the involvement scale asks respondents (CEOs or other top executives) to assess the extent to which board directors commit themselves sufficiently to fostering effective decisions and reversing failed policies. The mean score represents the board's involvement in the decision making of company (five items, α = 0.83). To establish measure validity, data about the functional and industry experience of board members, available for 87 firms in the sample, were collected from company web sites and annual reports. The average experience of board members correlates with the proposed survey-based measure (r = 0.68, p < .01), in support of construct validity.
For the measure of the level of board–CEO social ties, respondents considered the CEO's family or personal relationships with board members and indicated if board members are: (1) immediate family members; (2) acquaintances but not friends; and (3) friends. Excluding mere acquaintances provides a more precise measure of perceived friendship. In addition to the number of family members and perceived friends, the strength of such ties is measured as the closeness, duration, and frequency of each relationship (Granovetter, 1973). As is common for measuring social ties in organizational research (e.g., Brass, 1984; Perry-Smith, 2006), the product of the number of family members/perceived friends and tie strength is divided by the size of the board. This study considers both family relationships and friendship ties as one broad construct and sums the scores, because advice seeking literature does not argue any weight difference between them.
Market instability, characterized by the change rate of environmental factors relevant to strategic decision making (Duncan, 1972), is often a function of an industry's competitive dynamics. Following Wiersema and Bantel (1993), this study measures market instability as the changes in the industry concentration ratio, calculated as the percentage of an industry's sales accounted for by the four largest firms at the four-digit standard industrial classification level. Large absolute changes indicate high environmental instability, because they reflect shifts in market share because of new entrants, exits, consolidations, or erosion of market share, each of which is a clear sign of market instability. Data were collected from the TCISI database and the Taiwan Economic Journal (TEJ) database over a three-year period (2003–2005).
Board interlocks equal the total number of directorships held by all members of the board, divided by the number of board directors in the focal company (Zajac and Westphal, 1996). Data were collected from company prospectuses and annual reports over the three-year period prior to the survey.
The study also includes five control variables to minimize concerns about unobservable heterogeneity. Data were collected from the TEJ database. According to Anderson and Reeb (2003), family ownership is the fractional equity ownership held by the founder and his or her immediate family members. As a typical form of strong board–CEO bonding, the CEO duality measure involves coding sample firms in which the CEO is also the chairperson of the board as “1” and others as “0” (Rechner and Dalton, 1991). Firm size is a logarithm of the number of employees. The standard deviation of earnings per share serves as the measure of firm risk. The high-tech dummy is coded “1” if the firm-level R&D intensity (R&D expenses divided by sales) is greater than five per cent, and “0” otherwise. Several prior studies, such as Balkin, Markman and Gomez-Mejia (2000), also use this benchmark to identify high-technology firms. Data were first collected from the survey, and then validated by company proxy statements and board information disclosed in the TEJ database.
Moderated hierarchical regression analyses (Cohen and Cohen, 1983) are used to assess the effect of the independent variables on product innovation. The intercorrelations for all explanatory variables are examined using both bivariate correlations and variance inflation factors (VIF). The former show that the intercorrelations for all explanatory variables are less than 0.5 (see the descriptive statistics and correlation matrix in Table 1). The VIF analysis reveals no sign of multicollinearity, and the VIF values of all independent variables range between 1.37 and 4.45, far below the acceptable upper bound of 10. Both tests suggest that the regression estimates are not degraded by the presence of multicollinearity. Furthermore, the regression estimates consistently yield a Durbin–Watson statistic greater than 1.75, with a first-order correlation less than 0.20, which indicates autocorrelation is not a problem. Because the data consist of multiple cross-sections, heteroskedasticity might be an issue. Therefore, the residuals are plotted against firm size and the scale variable (Rajagopalan and Prescott, 1990); the results indicate an absence of heteroskedasticity in the data and, hence, no need for correction.
Descriptive Statistics and Correlation Matrix
| 1. NPI||3.09||0.64||1.00|| || || || || || || || || |
| 2. Family ownership||0.47||0.29||0.06||1.00|| || || || || || || || |
| 3. CEO duality||0.52||0.50||0.04||0.37||1.00|| || || || || || || |
| 4. Firm size||3.58||0.47||0.02||−0.05||−0.08||1.00|| || || || || || |
| 5. Firm risk||0.54||0.51||0.31||−0.01||−0.12||0.04||1.00|| || || || || |
| 6. High-tech dummy||0.59||0.39||0.33||−0.06||−0.03||0.00||0.35||1.00|| || || || |
| 7. Board involvement||3.51||0.44||0.13||0.35||0.02||0.03||0.18||0.13||1.00|| || || |
| 8. Social ties||2.51||0.74||0.10||0.28||−0.07||0.11||0.05||0.07||0.25||1.00|| || |
| 9. Market instability||2.83||0.75||0.43||0.02||−0.03||0.03||−0.04||0.00||−0.16||−0.17||1.00|| |
The results from the moderated hierarchical regression analyses appear in Table 2. The table reports standardized coefficients that indicate the effect that a one-standard-deviation change in an independent variable has on the outcome variable. Model 1 reports the baseline with only the control variables, which together explain 13 per cent of the variance (adjusted R2) in the data. In model 2, to test H1 and H3, two types of board/CEO relationships, in their linear and quadratic terms, are added to examine the non-monotonic effects of board involvement and board/CEO social ties. This model explains 29 per cent of the variance, and the change in variance explained is statistically significant (p < .001). A curvilinear relationship between board/CEO power balance and product innovation, as is borne out by the quadratic results (β = −1.74, p < .001), gives support to H1. Also, as predicted by H3, the board/CEO social ties are curvilinearly associated with the performance of new product introduction (β = −1.64, p < .01).
Results of the moderated hierarchical regression analysis
Dependent variable: the performance of new product introduction
|Family ownership|| ||0.12||1.66||0.10||1.58||0.09||1.46||0.09||1.61|
|CEO duality|| ||0.09||1.32||0.04||0.66||0.04||0.66||0.05||0.72|
|Firm size|| ||0.14||2.03*||0.13||1.85†||0.07||1.09||0.15||2.17*|
|Firm risk|| ||0.09||0.76||0.12||0.97||0.07||0.66||0.10||0.85|
|High-tech dummy|| ||0.26||2.09*||0.23||1.88†||0.28||2.56*||0.23||1.91†|
|Board involvement|| || || ||0.31||0.48||0.37||0.49||0.09||0.12|
|(Board involvement)2||H1|| || ||−1.74||−3.26***||−1.43||−2.87**||−2.42||−3.79***|
|Social ties|| || || ||0.39||0.55||0.47||0.70||0.31||0.43|
|(Social ties)2||H3|| || ||−1.64||−3.09**||−1.40||−2.82**||−2.28||−3.68***|
|Market instability|| || || || || ||0.21||2.43*|| || |
|Board interlocks|| || || || || || || ||0.11||1.26|
|Board involvement × Market instability||H2|| || || || ||−0.41||−6.59***|| || |
|Social ties × board interlocks||H4|| || || || || || ||0.28||2.85**|
| Adjusted R2|| || ||0.13|| ||0.29|| ||0.44|| ||0.33|
| Model F|| || ||5.75***|| ||4.89***|| ||5.23***|| ||4.90***|
| F for ▵R2|| || || || ||3.45**|| ||6.72***|| ||4.20*|
Models 3 and 4 are to test a contingency approach to the governance–innovation relationship. To avoid multicollinearity problems that might be introduced by multiple interaction terms, the independent variables are mean-centered to create the interaction terms. In Hypothesis 2, it is argued that environmental instability would negatively moderate the relationship between board/CEO power balance and firm innovativeness. The coefficient of the interaction of board involvement and environmental instability is negative and significant (β = −0.41, p < .001) in models 3, giving support to the hypothesis. For further confirmation, model 3, including the interaction term, explains 44.3 per cent of the variance. The change in explained variance relative to model 2 is also significant (p < .001). The results for model 4 show that the interaction effect of board/CEO social ties and board interlocks is also statistically significant (β = 0.28, p < .01), offering support to H4. The increase in explained variance of model 4 is again significant (p < .001), with 37.8 per cent of the variance explained by this model.
This study used slope analyses to confirm that the positive slope apparent at lower and mean level of and turns negative at higher level of dependent variable. The plots based on model 2 visually support the inverted U-shape of the innovation effects of two types of board/CEO relationships. The plots of model 3 and 4 likewise show the moderating effects of environmental instability and board interlocks on the governance–innovation relationship, wherein the level of environmental and interfirm characteristics experienced by a firm affects the slopes of the curves.
The support for H1 indicates that board involvement relates non-monotonically to new product introduction, and thus confirms conceptual discussions of board–CEO power balance that claim a power bias risks dysfunctional rivalries or fragmentation. Furthermore, this result suggests that a company board must maintain “an appropriate level of strategic involvement” to avoid encroaching on management's role. Complementarities based on a balanced division of power may engender more creative disagreement, and allow the upper echelons of the firm to attain more innovative ideas without fearing the loss of acceptance or damaged relationships.
In contrast to the sociopolitical force characterized by power balances in board–CEO relationships, this study considers how sociopsychological forces may influence board effectiveness with regard to new product decisions. The empirical results support H3: The provision of advice and counsel in the strategy-making process is more important for weaker board–CEO social ties, whereas their overdependence reduces board effectiveness. An inverted U-shaped relationship thus depicts the correlation between board–CEO social ties and product innovation performance. In turn, social ties induce CEOs to disclose information about organizational problems and seek advice from the board, which then reduces information asymmetry, and results in more informed involvement by the board. Consistent with the notion of leader–member exchange (Bauer and Green, 1996), this finding shows that a superior–subordinate relationship characterized by frequent advisory interactions (in addition to supervisory control interactions) can improve performance. However, too many personal connections may lead to extreme cohesion, and thus challenge the board's independence and objectiveness in relation to monitoring or disciplining managers. Therefore, these findings suggest that a mutually trusting board–CEO relationship complements rational controls, and makes board involvement more likely as a means of providing constructive feedback, as long as the social ties of both parties remain at an appropriate level.
By acknowledging that governance arrangements should adapt to changing conditions (Strebel, 2004), the contingency approach taken in this study advances our understanding of the context-dependent nature of corporate governance, because it addresses the situational efficacy of board–CEO relationships in different settings. According to the tests of H2 and H4, risk-alignment effects are conditioned by environmental and interfirm characteristics, respectively. The support for H2, which echoes Demb and Neubauer's (1992) argument, suggests that rising environmental ambiguity demands a paradoxical approach to governance. An unstable environment, which offers a context for greater managerial discretion, renders top executives more latitude over the course of innovative actions, and thus is disadvantageous to companies whose boards attempt to centralize their control over managerial decision making.
A firm's external network usually is considered a bridging approach for managing resource dependence (Meznar and Nigh, 1995), and interfirm linkages, such as board interlock ties, should serve as a lever for the advising functions of the board. Empirical support for H4 reveals that interlocking directorates positively moderate the innovation effects of board–CEO social ties; that is, directors’ ties to other boards increase their direct experience and indirect access to strategic information, which in turn enriches the knowledge structures they use to offer advice. Such linkages also strengthen the innovation-shaping effects of board–CEO social ties.
Limitations and Future Research Directions
This study enhances the body of knowledge pertaining to firm-level governance design, and the contexts that may nullify or strengthen the governance–innovation relationship. However, the results should be viewed in light of several limitations. First, this study considers CEOs as representatives of the top executive team. However, the board–management relationships may be more complex if other senior executives are considered. To fully capture the effects of cooperative strategic decision making on firm innovativeness, further research could examine the structural power of the entire executive team and their social ties with board members. Secondly, executives’ sociopsychological attributes could influence the board–CEO relationship. For example, some personalities may be more capable of remaining detached and objective while building trust between the board and top executives. Therefore, the effective mix of intrinsic/extrinsic motivation likely varies and, in turn, determines the performance implication of board–CEO relationships. In addition, a firm's situational characteristics, such as its status in the organization life cycle or its environmental munificence, may influence the ebb and flow of board–CEO relationships (Sundaramurthy and Lewis, 2003). Thirdly, in contrast with the cross-sectional nature of this study, a longitudinal study could better probe into the dynamics of governance paradoxes. However, because of the difficulties involved in collecting panel data, researchers may need to undertake creative uses of multiple data sources to facilitate research of this type. In addition to traditional time-series databases, they could use company documents, letters to shareholders, corporate speeches, and published histories to construct organizational histories. They also might quantify such historical data with content analysis, which would require coding the text to identify specific constructs and patterns. Fourthly, the empirical findings are based on Taiwanese firms. Although this study controls for two main features (i.e., family ownership and CEO duality) of Asian corporate landscape, the findings should be interpreted and generalized with caution, because board–CEO relationships in Taiwan may differ from those in other economic, institutional, and social environments. A broader sample of firms across national boundaries would make the empirical results more generalizable beyond the country-specific context provided herein.
This paper inquires why firms differ in terms of their new product performance, by demonstrating how formal and informal board–CEO relationships help determine a firm's entrepreneurial orientation. Both universal and contingency views were taken for this theme in a sample of Taiwanese firms. The former implies a general, direct link between two distinct types of board–CEO relationships and firm innovativeness, and thus documents the benefits and costs of governance properties across all contexts, ceteris paribus. In contrast, the contingency perspective provides a deeper look at environmental and interorganizational phenomena, and thus reveals situational contingencies and context-dependent prescriptions for governance practices.
Empirical support for the first set of findings (H1 and H3) suggests some important normative implications for governance design. Boards should fine tune their roles in the “oversight or micromanagement” dilemma, and thereby maximize the risk-alignment effect and enhance managerial risk preparedness. The innovation outcomes of two types of board–CEO relationships, not only echo Lewis's (2000) framework of organizational paradox, but also highlight the tensions in the distinct governance approaches. Many corporations apply formal logic based on internal consistency, and thus polarize governance prescriptions rather than stressing the interdependencies of board and top executive team. This distinction exacerbates the need to take sides, sparks defenses, impedes mutual learning, and fuels counterproductive reinforcing cycles. Managing the paradox of board–CEO relationships instead entails developing understanding and practices that accept and accommodate tensions.
The second set of findings (H2 and H4), based on a contingency perspective, also yields some context-dependent implications. Bourgeois and Eisenhardt (1988) argue that successful firms in high-velocity environments rely more on managerial intuition and quick responsiveness; this study confirms their observation by showing that the innovation effect of board dominance is negatively tempered by the instability of environment. Conditions in stable environments may be easier to control and factored into decisions, such that more board involvement and mechanistic monitoring of top executives is acceptable as a governance arrangement. In the steel or cement industry, for example, neither customers nor competitors are likely to change much, and industry standards have been well established; thus, discontinuous change is rare. In contrast, in the biopharmaceuticals or IT software industries, discontinuous changes occur frequently and shift the rules of the game (Dean and Shareman, 1996). In the face of high environmental instability, the potential that market conditions influence the viability of a strategic investment, such as new product introduction, increases and, thus, alters the causality between managerial power and decision-making effectiveness. Therefore, the optimal level of board dominance decreases in unstable environments compared with stable ones.
By recognizing that interfirm social network ties through directorate interlocks both facilitate knowledge exchanges and help manage resource dependence, this study clarifies a positive moderating role of directors’ social network in the form of their ties to other boards. The results likewise suggest that a board functions better as an effective advisor (or monitor) when its members also serve on the boards of other strategically relevant firms. Corporate leaders, therefore, should give considerable weight to the presence of the other board memberships of their director candidates. Board governance also should depart from a traditional insider/outsider distinction, common to the agency perspective, and embrace more finely tuned thinking that recognizes that board composition reflects the match between the external dependencies an organization faces and the resource acquisition potential of its board members.
Overall, this study enriches the body of knowledge in corporate governance by providing empirical evidence tested with robust data and methodology. The findings further have implications for two divergent types of innovation strategy, exploration, and exploitation. By definition, all innovation activities involve either the exploration of radically new ideas or the exploitation of existing routines (March, 1991). The former, as primarily investigated by this study, demands higher levels of managerial risk readiness. Acknowledging that good governance design softens managerial risk aversion, this study helps identify the governance conditions in which risk differentials between the board and CEO may align and, thus, when exploratory efforts, in the form of new product introduction, are more likely to succeed.