Institutionalized action and corporate governance: the reliance on rules of CEO succession.
This paper explores the reliance on formal and informal rules in corporate governance. The principal idea is that corporate boards of directors are guided in their decisions by both historical precedents and formal rules that are not easily modified as board actions become institutionalized and norms of appropriate beliefs and behavior become established (Zucker, 1977; March and Olsen, 1989). Building on an institutional theory of action (March and Olsen, 1989), it presents a new explanation for inertia in corporate governance and shows that CEO succession is a routinized process, conditioned by formal and informal rules. The rules and routines of executive succession and selection embody the board's role in what Weber (1978) called the "routinization of charisma" as the power and authority of the founding chief executive becomes depersonalized and is invested in the formal position rather than in the individual occupant.
This paper contributes both to the theoretical literatures on corporate governance and organizational rules and to the empirical research on CEO succession and selection. The theory focuses on the consequences of rules in structuring organizational action and decision making, while the empirical research examines the reliance on formal and informal rules and routines of insider versus outsider CEO succession in U.S. industrial corporations from 1960 to 1990. Unlike much previous work on the selection of insiders versus outsiders that examines CEO selection conditional on succession having occurred (Boeker and Goodstein, 1993; Cannella and Lubatkin, 1993), this study examines CEO succession and selection as interdependent. An emphasis on how the reliance on rules of CEO succession persists despite changes in the firm's performance and other moderating influences distinguishes this paper from adaptationist perspectives on corporate governance, which imply that executive succession is unlikely to be routinized.
An institutional theory of action (March and Olsen, 1989; March, 1994) highlights how rules enable and constrain organizational decisions and those of its boards of directors. This theory is part of a broader institutionalist perspective in contemporary social and organization theory (Scott, 1995). Three aspects of the theory are important in explaining corporate governance. First, it proposes a rule-based theory of action and decision making, in which organizational decisions are enacted through the application of appropriate rules (March, 1994). It distinguishes this rule-based theory of action from theories of bounded rationality or interest-based approaches to rule following. Second, it highlights the importance of history in shaping the rules and routines that structure behavior (March and Olsen, 1989). Third, and building on the theoretical foundations on an earlier Carnegie school tradition (March and Simon, 1958; Cyert and March, 1963), it emphasizes the interplay of cognitive and political factors in the generation and maintenance of organizational rules.(1) Compared with other versions of institutionalism (e.g., Meyer and Rowan, 1977; DiMaggio and Powell, 1983), this perspective emphasizes organizational-level rules versus the effects of the broader institutional environment, focuses on explaining actions and decisions rather than structures, and provides an explicit alternative to the theory of rational choice. In rational theories of choice, decisions are based on the evaluation of alternatives in terms of their consequences for preferences or values of the decision maker. An institutional theory of action (March and Olsen, 1989) provides an alternative, "a logic of appropriateness by which actions are matched to situations by means of rules organized into identities. . . . When individuals and organizations fulfill identities, they follow rules or procedures they see as appropriate to the situation in which they find themselves" (March, 1994: 57).
In this paper, I follow Giddens (1984: 21) in defining rules as history-dependent, socially constructed techniques, or generalizable procedures applied in the enactment and reproduction of social structure. They include both the formally codified standards and the informal routines and norms that structure conduct in organizations. Rules are embodied in the policies, programs, procedures, routines, and conventions around which organizational activities are constructed. They comprise the knowledge, capabilities, beliefs, values, and memory of the organization and its members and are invoked in response to internal and external stimuli. Rules are not entirely derived from social relationships or other elements of social structure but have independent effects on social phenomena. Rules do not serve any blind functional necessity but are tools that both empower and control the social construction of organizational practices. They are both a source of organizational inertia (March and Olsen, 1989) and a guide for organizational adaptation and change (Ocasio, 1995).
Contrary to some neoinstitutionalist perspectives that focus on cognitive foundations of rules at the expense of political factors (DiMaggio and Powell, 1991), this paper emphasizes both.(2) In the absence of agreed-upon rules and guiding principles, interests have critical import in the social construction of rules. Furthermore, reliance on rules does not imply a lack of political action and behavior; instead, rules establish the parameters by which the political game is played. In the context of corporate governance and chief executive officer (CEO) succession, the rules of insider succession that prevail in most organizations establish the guidelines for internal political dynamics and contests for control among internal contenders for the CEO's position (Ocasio, 1994). Nevertheless, what distinguishes an institutional theory of action from interest-based approaches is that according to this view, while rules shape and are shaped by political processes, they are not merely the reflection of the interests of decision makers but are shaped by history and experience and are not easily changed in response to the immediate interests of organizational decision makers.
Rules of corporate governance pervade all forms of decision making by boards of directors. Examples of formal rules of corporate governance include reliance on committees of outside directors for setting executive compensation and for monitoring corporate audits, approval by boards of all capital expenditure decisions above a certain amount, and standardization of the CEO succession and selection process. Examples of informal rules of corporate governance include limits on open criticism of the CEO, not contacting fellow board members outside of meetings, and, for most boards, using a rhetoric of shareholder interests to explain board decisions (Mace, 1971; Lorsch, 1989).
Formalization of rules and bureaucratic control. The separation of ownership and control and the rise of managerial capitalism led to the rise of bureaucratic forms of organizations in U.S. industrial organizations (Bede and Means, 1967; Chandler, 1977). A key identifying characteristic of corporate bureaucracies is the degree of formalization, or the reliance on formal rules by which social positions and practices are defined independently of the personal characteristics and values of organizational participants. An institutionalized action perspective highlights the role of corporate boards of directors as institutional agents in enacting and reproducing formal rules and bureaucratic controls. In corporate bureaucracies, the sources of control by corporate directors reside not in their economic or social capital but in the formal authority derived from their position. Formal rules thereby provide a legitimate basis for actions and decisions by boards of directors and allow for the reproduction of the legitimate basis of authority (Weber, 1978). In addition, the formalization of practices allows for a stable set of expectations about the decisions of other board members and makes behavior more predictable by standardizing and regulating action (Simon, 1997). Formalization also serves to objectify the rules, making the definitions of rules and objectives appear objective and external to organizational actors (Zucker, 1977).
Enabling function of rules. While theories of rational decision making often highlight how rules serve to constrain action and behavior, rules also empower and facilitate action, increasing the likelihood that organizational decisions will be made rather than postponed. First, rules provide readily available solutions, scripts, and procedures for the enactment of organizational decisions (Cohen, March, and Olsen, 1972; March and Olsen, 1976). Given bounded rationality, the existence of formal and informal rules in organizational memory decreases the need for search and provides readily available routines or programs of action that structure behavior (March and Simon, 1993). Given the ambiguity inherent in corporate governance, the reliance on rules facilitates action by presenting board members with readily available solutions for organizational problems. Second, reliance on rules provides for accountability and reliability in organizational activities (Hannan and Freeman, 1984) and enhances the ability of board members to meet their fiduciary responsibilities. Corporate boards of directors are accountable for their actions and decisions. Reliance on the prevailing rules increases the board's ability to justify and defend its actions and decisions. Finally, rules help enable board actions and decisions by limiting and channeling political conflict in organizations. Rules and routines serve as a form of political truce (Nelson and Winter, 1982), with stable accommodations and quasi-resolution of conflict among the divergent interests in the organization (Cyert and March, 1963). Reliance on rules and routinized procedures of corporate governance signals the political viability and appropriateness of board decisions. In the absence of rule-based action, political mobilization is required, with the resultant problems of conflict and collective action and a decreased likelihood that decisions will be made and implemented.
Rules and common social identification by board members. An institutional theory of action posits that decisions are made by matching situations to rules according to the identities of the decision makers (March and Olsen, 1989). Decisions follow a logic of appropriateness in which action results from the application of rules to the question, "What does a person such as I, or an organization such as this, do in a situation such as this?" (March, 1994: 58). The logic of appropriateness that guides the board's behavior is shaped by both the legal fiduciary requirements (Clark, 1991) and the normative commitments that board members make to exercise the duties of loyalty, care, due diligence, and good business judgment, with their ultimate responsibility to "do the right thing" (Lorsch, 1989: 70). Rules become infused with value, as symbols of the organizational mission and as embodiments of institutional purpose (Selznick, 1957), increasing social identification among the organization's executives and its board of directors.
Reliance on rules and routines is not automatic but is conditional on decision makers identifying with the available rules (Gouldner, 1954), which presupposes that board members share a common social identification. Both past quantitative and qualitative empirical research on corporate governance supports the view that board members share a common identification resulting from shared values, commitments, and demographic characteristics. Westphal and Zajac (1995) found that common demographic characteristics were a principal predictor of new director selection and that social similarity shaped boards' decisions on executive compensation. Hirsch (1982: 12), in interviews of board members of Fortune 500 companies, found that directors do not think of themselves as representing any specific interest or organizations in their role as board members but are self-regulated by a "widely shared normative structure of appropriate behavior and forms of control." Useem (1984) found similar coherence, commonalties, and a generalized consciousness among corporate boards of directors in large corporations in the United States and the United Kingdom. This coherence and commonality of purpose and identification among board members fosters their reliance on rules of appropriate behavior.
Reliance on rules and inertia in corporate governance. A rule-based enactment of organizational decision making implies that the reliance on rules of corporate governance persists as boards' actions become routinized and norms of appropriate beliefs and behavior are established (Zucker, 1977; March and Olsen, 1989). According to this view, reliance on rules will continue under a variety of organizational contingencies, including poor economic performance and involuntary CEO departures. The reliance on rules of board decision making and the persistence of their effects implies that the formal mechanisms of corporate governance are characterized by organizational inertia (Hannah and Freeman, 1984).
The persistence of the effects of rules of corporate governance under changing organizational contingencies distinguishes an institutional theory of action from theories that are based on bounded rationality. Adaptationist theories imply that board rules and routines, if observed, are an ephiphenomenon, a reflection of bounded rationality in the pursuit of material interests by the board's dominant coalition (Cyert and March, 1963; Pfeffer and Salancik, 1978). In particular, the logic of consequence and interests that underlies adaptationist theories implies that board rules are unlikely to persist as the conditions that led to their establishment vary, as failures in performance trigger a search for new rules, and as the material interests of controlling board members change. While the persistence of board rules under poor economic performance is not easily reconciled with adaptationist views on corporate governance, it is an observable implication of an institutional perspective on action. Therefore, an examination of how the rules of corporate governance are moderated by performance and other organizational contingencies allows us to test empirically whether rules best reflect either a logic of consequences (March and Olsen, 1989), with bounded rationality in pursuit of interest and adaptation, or decision making following a logic of appropriateness (March and Olsen, 1989).
Rules of CEO succession. The succession of an insider versus an outsider as CEO provides an important context for applying an institutionalized action perspective on corporate governance. The CEO succession process, typically viewed as a mechanism for organizational adaptation (Pfeffer and Salancik, 1978; Tushman and Romanelli, 1985), is guided by the formal and informal rules of corporate governance. Outsider succession, in particular, has been seen as a trigger of organizational change (Helmich and Brown, 1972). While boards may delegate to CEOs as agents important decisions that occur within their tenure, the succession of a new CEO provides an opportunity for realigning the corporation with the controlling interests of the board of directors. In particular, both agency and managerialist accounts of corporate governance highlight the importance of CEO succession as a mechanism for adapting the organization to environmental and economic contingencies (e.g., Pfeffer and Salancik, 1978; Zajac, 1990). These accounts suggest that CEO succession is unlikely to follow routinized procedures but will reflect, instead, the interests of those who control the corporation. But these accounts also imply more calculative action than is likely to occur given the ambiguity created by a CEO succession.
Managerial succession is a potentially disruptive activity for organizations (Gouldner, 1954; Grusky, 1961; Carroll, 1984). At the same time, the effects of succession on corporate performance are highly ambiguous, and even the results of academic research on whether it is likely to increase or decrease firm profitability or its survival prospects have been equivocal (e.g., Carroll, 1984; Zajac, 1990). Given the potential hazards and ambiguity surrounding the CEO succession and selection process, boards are likely to focus on controlling the accountability and reliability of organizational activities. Boards' reliance on the rules and routines of executive succession and selection provides a sense of order in governance decisions, communicates commitment to the corporate mission and programs, and channels political dynamics into appropriate practices and procedures. According to this perspective, organizational change is mediated through the existing rules and institutional logic (Ocasio, 1995). Boards of directors rely on the rules of CEO succession to implement change and to guide executive turnover.
According to the institutional theory of action (March and Olsen, 1989), an important, albeit frequently ignored aspect of rules is their enabling function in organizational decision making. Rules reduce ambiguity and provide a set of acceptable solutions that can be invoked in organizational problem solving and decision making. The institutional theory of action provides a cognitive perspective on decision making consistent with garbage can models and the theory of ambiguity and choice (Cohen, March, and Olsen, 1972; March and Olsen, 1976; DiMaggio and Powell, 1991). According to this theory, the likelihood of a decision being made depends on the repertoire of readily available solutions. For CEO succession, the likelihood of a succession event being enacted by the corporate board of directors depends on the availability of succession as a solution. The existence of specific rules of succession increases the availability of succession as a solution to the problems of the corporation and its board of directors, thereby increasing the rate of executive succession. Furthermore, the rules of insider and outsider succession can have different effects on the likelihood of insider versus outsider succession.
Formal rules of insider succession. Formal rules and procedures have particular importance for the process of executive succession and selection. The depersonalization of power in corporate bureaucracies requires that succession be guided by formal rules and procedures. An important aspect of this formalization is the creation and reliance on internal labor markets (ILMs) for top management executives, including the CEO. Reliance on internal labor markets for managerial succession is a central institutional characteristic of managerial capitalism in U.S. industrial corporations (Chandler, 1977). While any succession process entails a contest for power among potential candidates for the executive position (Ocasio, 1994), the vesting of formal authority is facilitated when executive transitions follow a bureaucratic order and formalized process. CEO succession and selection become part of the bureaucratic routine as the depersonalization of the formal authority of the chief executive results in what Weber (1978) termed the "routinization of charisma." Through the application of the formal rules of succession, the prevailing sources of power shift from the personal power and charisma of the founding executive to the formal authority embodied in the rules of corporate governance.
Vancil (1987), in his descriptive study of the succession process in large U.S. corporations, found evidence for the prevalence of two distinct variants of formal rules of insider CEO succession. The more common form was the existence of an heir apparent in the form of a president and chief operating officer (COO), distinct from the CEO. Cannella and Lubatkin (1993), following a managerialist theory interpretation of Vancil's study, found that the designation of an heir apparent decoupled the relationship between firm performance and outsider succession. The second variant was the reliance on a "horse race" among internal competitors for the top spot, usually termed as vice chairman or executive vice presidents. In both variants, the contenders for the CEO position were also members of the board of directors, which afforded on-the-job training for the candidates and provided outside board members with the opportunity to gain knowledge and experience about the inside contenders.
Both types of insider CEO succession constitute a formalization of an internal labor market for CEOs, which embeds the rules of executive succession within the formal authority structure of the corporation. This internal market for CEOs is characterized, as are other forms, by promotion from within, the salience of formal job titles and career ladders, and a reliance on and rigidity of formal rules and procedures for decision making (Doeringer and Piore, 1971). Boards will commit themselves to insider CEO candidates who have been formally designated by both their job title and board membership as either heirs apparent (Cannella and Lubatkin, 1993) or as potential replacements for the CEO. Vancil's study suggests that the prevalence of formalized procedures for executive succession increases the likelihood that insider succession will take place. When faced with formalized internal labor markets for CEOs, the logic of appropriateness and rule-based enactment of decisions by board members will lead them to rely on prevailing succession routines and increase the likelihood of insider CEO succession. Furthermore, the formalization of the executive succession process provides the board with readily available candidates for the CEO's position and rivals for the top executive's power:
Hypothesis 1 (H1): The formal designation of executives (other than the CEO) as directors and as president, chief operating officer, or vice chairman increases the likelihood of insider CEO succession for incumbent CEOs.(3)
Precedents as informal rules of CEO succession. A primary source for the generation of rules is the organization's historical experience (Selznick, 1957). History and precedents shape organizational cognition, particularly under conditions of ambiguity (Zucker, 1977). The existence of a historical precedent structures future possibilities for the enactment of organizational decisions. With the adoption of a program of action, it becomes embedded in the memory of the organization and its decision makers. Consequently, organizations will be more likely to adopt programs of action to the extent that they follow existing precedents and have been adopted in the past (March and Olsen, 1989).
Kelly and Amburgey (1991) found evidence of the effects of past precedents of organizational adoption in the airline industry, with organizations significantly more likely to repeat changes they had experienced in the past, extrapolating past trends and applying them to new forms of environmental change. This reliance on history and past precedent for decisions of boards of directors was also found by Bjorkman (1989) for foreign direct investments and Amburgey and Miner (1992) for corporate mergers. When the CEO succession involves a founder, however, no historical precedents are available.
According to the theory, the selection of succession as a program of action will depend on the availability of solutions, on their appropriateness, and on the degree of political support and commitment to the prevailing rules. In the absence of precedents, solutions are not readily available, no rules of appropriate behavior have been established, and no rules have gained political commitment or support. In the case of founders, both cognitive and political factors hinder boards' ability to enact either insider or outsider succession as a response to the environmental contingencies faced by the firm. In the absence of experience with CEO succession, the possibility that the firm can surmount the disruptive effects of succession is not known, and a change in executive is not readily available in the organization's repertoire of action possibilities. For founders, the power of the executive is more closely tied to his or her personal power and charismatic authority, rather than to the authority invested in the position. The enabling function of historical precedents on insider and outsider succession implies that founders will have lower rates of both insider and outsider succession because of the lack of historical experience:
Hypothesis 2a (H2a): Founders will have lower rates of both insider and outsider CEO succession.(4)
With the succession of the founding CEO, succession becomes part of the firm's and the board's repertoire of action possibilities, and the charismatic basis of the executive's power shifts to a bureaucratic basis, embodied in the rules and routines of CEO succession (Weber, 1978). The cognitive and political factors that influence the reliance on informal rules of succession may vary in strength, however, for either insider or outsider CEO succession. These effects must be understood in light of the broader institutional field of corporate governance in which publicly held corporations operate, where the institutionalization of action occurs not only at the level of the individual organization but at the level of the organizational field (DiMaggio and Powell, 1983). Statistically, insider succession has become the norm for U.S. publicly held corporations, particularly for larger companies, though firm-level effects are likely to matter, since experience with insider and outsider succession will alter boards' decisions through the routinization of charismatic authority and the separation of the firm's identity from that of its founders.
Precedents of insider versus outsider succession may have different effects on their availability as solutions for the boards of directors to consider, according to their appropriateness and their level of support by the reigning political coalition. First, past precedents of outsider succession, relative to precedents of insider succession, increase their availability in the board's repertoire of activities and signal their political appropriateness in the focal organization. Given that insider succession is the statistical norm for the industrial sector, insider succession is likely to be a more available solution than outsider succession. The political effects of insider CEO succession are likely to be greater than those of outsider succession, nonetheless, as informal commitments and implicit contracts between the board and top executives parallels the effects of more formal internal labor markets for CEOs. For precedents of outsider succession, there are no similar individuals or group of individuals that sustain an implicit contract with the board. The different effects of reliance on the prevailing rules and precedents of insider versus outsider CEO succession warrant separate consideration of the effects of firm-level insider and outsider CEO experience on subsequent CEO succession:
Hypothesis 2b (H2b): Past reliance on insider succession increases the likelihood of insider CEO succession for incumbent CEOs.
Hypothesis 2c (H2c): Past reliance on outsider succession increases the likelihood of outsider CEO succession for incumbent CEOs.
While adaptationist views on corporate governance could account for the reliance on rules of insider succession based on either the superior average performance of insiders as CEOs (Zajac, 1990) or the power of insiders over the board of directors (Cannella and Lubatkin, 1993), they can less directly account for the reliance on rules of outsider succession. H2c is thereby particularly important in testing the applicability of an institutionalized theory of action to corporate governance and in contrasting its predictions with those of alternative theories. In a managerialist account, outside succession is not a preferred alternative and is unlikely to persist.
Firm age and the reliance on rules. Earlier empirical work from an institutional theory of action perspective has found that the reliance on rules increases with the age of the organization (Zhou, 1993). The mission and purpose of the organization becomes embodied in the norms and rules of the organization, and organizational rules become taken for granted and less subject to question or challenge. Hannan and Freeman (1984, 1989) argued that organizational histories generate constraint on structural change by providing legitimate justifications beyond self-interest for opposing change and precluding consideration of alternatives. Consequently, the reproducibility of organizational rules increases monotonically with age. Furthermore, pressures toward internal consistency and homogeneity of members' perceptions also suggest that organizational rules are likely to increase with age (Aldrich and Auster, 1986). These arguments suggest a direct relationship between firm age and the reliance on rules of CEO succession and the selection of insiders and outsiders as CEOs:
Hypothesis 3a (H3a): Firm age will interact with the formal designation of executives (other than the CEO) as directors and as president, chief operating officer, or vice chairman to increase the rate of insider succession for incumbent CEOs.
Hypothesis 3b (H3b): Firm age will interact with the prevalence of informal precedents of insider succession to increase the rate of insider succession for incumbent CEOs.
Hypothesis 3c (H3c): Firm age will interact with the prevalence of informal precedents of outsider succession to increase the rate of outsider succession for incumbent CEOs.
Effects of prevalence of formal rules. The effect of rules may increase with their use, and there may be moderating effects of past prevalence of formal internal labor markets for CEOs on the rate of insider succession. The effects of formal rules are more likely to become taken for granted, the greater their past usage in the firm (Zucker, 1977). The greater the prevalence of formal rules, the more these rules will become institutionalized and the greater their effect:
Hypothesis 4 (H4): The proportion of years in which a firm has formally designated executives (other than the CEO) as directors and as president, chief operating officer, or vice chairman will increase the rate of insider succession for incumbent CEOs.
Changing institutional context and the reliance on rules. The institutionalized action perspective (March and Olsen, 1989) used in this paper focuses on the effects of organizational-level rules on the rates of insider and outsider succession. An important, albeit relatively unexplored question remains about the relationship between organizational-level rules and rules prevailing in the broader institutional environment (Meyer and Rowan, 1977; DiMaggio and Powell, 1983). While the firms in this study are all part of the U.S. industrial sector, the rules of corporate governance have changed during the time period studied, particularly during the 1980s (Useem, 1996). For the industrial sector as a whole, insider succession was the statistical norm, with increasing rates of outsider succession during the period.
DiMaggio and Powell (1983) argued that the strength of isomorphic forces depends on the centralization of resources and the extent of professionalization in an organizational field. Given the decentralized nature of resource allocation and the variance in the professional and functional backgrounds of executives in the U.S. industrial sector (Fligstein, 1990), the strength of isomorphism in the U.S. industrial sector is likely to be lower than in other sectors, such as health care or education. While this relatively low degree of centralization and professionalization allows for organizational-level variation in the rules of CEO succession, mimetic isomorphic processes are still likely to be prevalent. Reliance on formal and informal organizational rules of succession is likely to depend on the relative prevalence of insider and outsider succession in the sector. While insider succession remains the statistical norm for the sector, outsider succession increased for the sample relative to insider succession during the period studied. As isomorphic change (DiMaggio and Powell, 1983) increases the reliance at the sectoral level on outsider succession, the appropriateness of insider succession is likely to decrease relative to outsider succession. Therefore, the reliance on rules of formal ILMs and past precedents of insider succession is likely to decrease over time, with a parallel increase in the reliance on informal precedents of outsider succession.
To explore the effect of changes in the broader institutional context on the reliance on organizational-level rules, the moderating effects of historical time need to be examined. Previous research by Ocasio (1994) has shown that historical effects on succession are best captured through the year of hire of the incumbent CEO. This implies that CEOs are affected by changing rules and expectations about CEO succession that were prevalent in the broader institutional context during the beginning of their tenure and that these rules are imprinted for the remainder of their tenure. Therefore:
Hypothesis 5a (H5a): The year of hire of the CEO will interact with the formal designation of executives (other than the CEO) as directors and as president, chief operating officer, or vice chairman to decrease the rate of insider succession for incumbent CEOs.
Hypothesis 5b (H5b): The year of hire of the CEO will interact with the prevalence of informal precedents of insider succession to decrease the rate of insider succession for incumbent CEOs.
Hypothesis 5c (H5c): The year of hire of the CEO will interact with the prevalence of informal precedents of outsider succession to increase the rate of outsider succession for incumbent CEOs.
Alternative Explanations for Reliance on Rules
While an institutional theory of action provides a unified, parsimonious account of the previous hypotheses, it is important to show that the reliance on the rules of CEO succession cannot be readily explained by alternative accounts.
Several perspectives provide rival interpretations of one or more of the hypotheses presented. In particular, arguments based on bounded rationality, the founder's power, and board structure provide alternative accounts for the use of rules. Unlike an institutional theory of action, in which rules are persistent and not readily amenable to change, however, these alternative explanations suggest that reliance on rules will most likely occur under certain conditions but not others.
Bounded rationality. According to theories of bounded rationality (Simon, 1957; March and Simon, 1958; Cyert and March, 1963), reliance on rules, programs, and routines is contingent on favorable economic performance, and organizations will undertake failure-induced changes in rules when performance is poor. The effects of economic performance on persistence and change in the rules of CEO succession thereby provide a test of whether theories of bounded rationality can better account for rule following than the logic of appropriateness and an institutional theory of action. If theories of bounded rationality account for CEO succession rules, reliance on rules of CEO succession should vary with the economic performance of the firm and should have subsequent effects on insider and outsider succession. Thus, there should be significant interaction effects between firm performance and the formal and informal rules of CEO succession:
Hypothesis 6a (H6a): Firm performance will interact with the formal designation of executives (other than the CEO) as directors and as president, chief operating officer, or vice chairman to increase the rate of insider CEO succession for incumbent CEOs.
Hypothesis 6b (H6b): Firm performance will interact with the prevalence of informal precedents of insider succession to increase the rate of insider succession for incumbent CEOs.
Hypothesis 6c (H6c): Firm performance will interact with the prevalence of informal precedents of outsider succession to increase the rate of outsider succession for incumbent CEOs.
In contrast to theories of bounded rationality, an institutional theory of action implies that commitment to the board's values, norms, and beliefs will continue under poor economic performance. The rules of corporate governance provide direction and guidance for the board's decision making both when performance is poor and when performance is favorable. Boards respond to economic adversity, not by evaluating the economic and political consequences of whether they should persist with their past choices or whether they should change, but by determining what is the appropriate thing to do as determined by the prevailing rules and routines of the organization, including those of CEO succession.
Founder power. Following an institutional theory of action, H2a stated that founders would have lower rates of both insider and outsider succession than other CEOs. This effect has been found in previous empirical research and has been attributed to the power of the founder (McEachern, 1975). An institutional theory of action does not imply that the founder's power is unimportant but, instead, that the absence of rules of CEO succession increases the power of founders and decreases the likelihood that succession will take place. In addition to the power of executives based on rules or their absence, the power of founders may increase from their social and economic capital, as measured by founder tenure. To discount alternative explanations that attribute the increased power of the founder to social and economic capital, rather than to rule following, this study tested whether founder tenure decreases the rate of insider and outsider succession, with the following hypothesis:
Hypothesis 7 (H7): Founder tenure decreases the rate of insider and outsider CEO succession.
Effects of early CEO departures versus retirements. Adaptationist theories of corporate governance and executive succession often distinguish between the effects of CEO dismissals and early departures and other forms of succession, particularly retirements (Fredrickson, Hambrick, and Baumrin, 1988). Empirical studies have supported this view, finding that the determinants of CEO succession and selection vary for CEOs facing retirement (i e., 64 years and older) and younger executives (Puffer and Weintrop, 1991; Cannella and Lubatkin, 1993). This adaptationist interpretation implies that the reliance on rules of CEO succession is more likely to hold for the normal process of retirement. For CEO dismissals and early departures, corporate governance is best understood as shaped by economic performance and political processes, and boards are unlikely to rely on rules of CEO succession. Consequently, according to this perspective, the reliance on rules of CEO succession will hold only for CEOs facing retirement:
Hypothesis 8a (H8a): The formal designation of executives (other than the CEO) as directors and as president, chief operating officer, or vice chairman will increase the rate of insider CEO succession for incumbent CEOs facing retirement age, but not for younger executives.
Hypothesis 8b (H8b): Informal precedents of insider succession will increase the rate of insider succession for incumbent CEOs facing retirement age, but not for younger executives.
Hypothesis 8c (H8c): Informal precedents of outsider succession will increase the rate of outsider succession for incumbent CEOs facing retirement age, but not for younger executives.
Alternatively, the institutional action perspective on corporate governance implies that boards will respond to the contingency of early CEO departures by relying on prevailing rules. Consequently, an institutionalized action perspective suggests that the formal and informal rules of CEO succession apply to cases of early CEO departures as well as to retirements.
Outside directors and board structure. Both agency theory (Jensen and Meckling, 1976; Weisbach, 1988) and structural perspectives on board independence (Westphal and Zajac, 1995) emphasize the relative roles of inside and outside directors in shaping the outcomes of corporate governance. According to these perspectives, given internal labor markets for CEOs, the greater selection of inside directors may result not from rule following and the logic of appropriateness but from either the rational, adaptive strategy of outside directors or their social ties with prospective candidates? Both arguments imply that the effects of internal labor markets on insider succession will be greater when there is a greater proportion of outside than inside directors on the board:
Hypothesis 9 (H9): The proportion of outside directors on the board will interact with the formal designation of executives (other than the CEO) as directors and as president, chief operating officer, or vice chairman to increase the likelihood of insider CEO succession for incumbent CEOs.
A random sample of 120 U.S. industrial corporations, or 4.45 percent of the total population listed in the Moody's Industrial Directory for 1980, was selected for the analysis. This sample includes publicly held companies listed in the New York, American, or regional stock exchanges and contains a broad representation of firm sizes and ownership structures. This sample differs substantially from Fortune 500 firms in both size and the prevalence of insider and outsider succession. The mean size of the firms in the study is approximately 3,000 employees. Past studies of Fortune 500 firms have shown that outsider selection occurs in approximately 5 percent of all cases. In the current sample, the prevalence of outsider selection is approximately 25 percent. For testing the hypotheses, this sample is preferable to Fortune 500 firms because it provides greater variance in the independent variables, with a larger proportion of firms having past precedents of outsider succession. In Fortune 500 firms, formal internal labor markets and past precedents for insiders are dominant, with very limited variance on the prevailing rules of CEO succession.
The unit of observation is the company year, covering the years 1960-1990. I dropped from the sample 14 firms for which financial data were missing, leaving 108 firms. Not all companies had data for the entire period. Many were founded and/or became publicly held after 1960. Many others merged, became bankrupt, went private, or otherwise ceased to be publicly held companies during the 1980s. I treated any spells ending in bankruptcy, acquisition, and change to private ownership as right-censored at that point. This implies that the effects being measured relate exclusively to normal forms of succession within the current ownership of the firm. The final sample included 2,286 company-years of data.
The sample was selected as of 1980 to permit corporations founded since 1960, including high-technology companies, to become part of the sample. This creates some potential for sample selection bias, as firms that disappeared between 1960 and 1980 were excluded from the sample. Sampling as of 1960, however, would have excluded newer firms from the sample, which then would have been less representative of firms in 1990. Sampling in 1980 was a compromise solution that would both reduce sample selection bias and produce a representative sample of industrial firms in 1990.
Independent Variables and CEO Succession Events
I coded CEO succession events from Standard and Poor's Directory of Corporations, Officers, and Directors based on changes in the names of the relevant officers. A total of 216 succession events occurred in the sample. Outsiders are sometimes appointed to the corporation for a short period of time in preparation for becoming CEOs. Therefore, following Cannella and Lubatkin (1993), CEOs were coded as insiders if they had been employees of the company for at least two years prior to becoming CEO; otherwise they were classified as outsiders. Data from Standard and Poor's Directory of Corporations, Officers, and Directors was supplemented by proxy statements, 10Ks, annual reports, and information from Who's Who in Industry and Finance to verify prior employment history. Outside members of the board of directors who were appointed as CEOs were classified as outsiders. A total of 157 insider succession events and 59 outsider successions were recorded.
Formal Rules of CEO Succession
I coded the existence of formal rules, or internal labor markets, for insider CEO succession based on the job titles of inside directors other than the CEO. Job titles provide institutional accounts of the roles, functions, responsibilities, and expectations of the holders of the formally designated jobs, as well as their relative status, ranking, and position within the internal labor market for executives. I recorded a formal ILM when a president, chief operating officer, or vice chairman (other than the CEO) was listed as both an officer of the company and a member of the board in Standard and Poor's Directory of Corporations, Officers, and Directors. Job titles of president, chief operating officer, and vice chairman (other than the CEO), coupled with the individuals' designation as inside directors, signal to the internal and external communities that their occupants are potential contenders for the chief executive position (Vancil, 1976). Their designation in these positions thereby signals the existence of formal rules and procedures for executive succession. Formal ILMs for insider succession were in place in 65.4 percent of the CEO-years in the sample.
Informal Rules of CEO Succession
I measured reliance on informal rules of CEO succession by examining whether the firm had relied exclusively in the past (since 1950) on either insiders or outsiders or whether it had experienced both. I recorded the history of all CEOs in the company since 1950 from Standard and Poor's Directory of Corporations, Officers, and Directors. CEOs of new companies were classified as founders, others as insiders or outsiders, using the coding scheme and data sources described above. I coded four mutually exclusive categories of past succession experience: founder, when the current CEO was the founder, past insiders, when all CEOs since 1950, other than the founder, were insiders, past outsiders, when all CEOs since 1950, other than the founder, were outsiders, and both, when CEOs since 1950 included both insiders and outsiders. In the sample 52.0 percent of the company-years were classified as past insiders, 12.6 percent as past outsiders, 15.3 percent as both insiders and outsiders, and 20.1 percent as founders.
Moderators of the Reliance on Rules
Firm age was calculated from the date given in Standard and Poor's Directory of Corporations, Officers, and Directors of the year of the firm's initial founding or incorporation. Data on firm ownership were obtained from Value Line Investment Survey and from corporate proxy statements. Proportion of years with ILM was constructed by taking the past history of formal designation of internal labor markets and calculating the number of years that a firm had an ILM for CEOs as a proportion of total years the firm had been in the sample, up to and including the current year. This variable was used as a measure of the prevalence of formal ILMs for CEOs. It is subject to some error in measurement because the years prior to 1960 or when the firm was privately held are not included in the proportion. Year of hire measured the year the CEO was hired (minus 1900) and was intended to capture historical trends in the rate of CEO succession. I expected a positive trend for the main effect, taking into account increased pressures on CEOs during the 1960-1990 period studied, particularly for outsider succession. The variable was also used to measure the effects of changing rules of insider succession in the industrial sector. The assumption in this variable is that CEOs are affected by changing rules and expectations about CEO succession that were prevalent during the beginning of their tenure and that these rules were set for the remainder of their tenure. This variable provides a much better fit than an alternative measure of time trend, in which it is assumed that changing rules affect new and incumbent CEOs equally.
Performance. I used return on assets (ROA), adjusted for industry average, obtained from COMPUSTAT as an economic performance measure, given past studies of CEO succession and selection and Gibbons and Murphy's (1990) finding that relative ROA affects CEO compensation. The correlation between adjusted and unadjusted ROA was .94, and the results do not change materially if unadjusted ROA is used as a measure of performance. Data on ROA were lagged one year. Two- and five-year averages were also estimated, but a one-year lag increased the variance explained by the model. Following Puffer and Weintrop (1991), I classified retirement age as 64 years old or over. Data on the CEO's age came from Standard and Poor's Directory of Corporations, Officers, and Directors, proxy statements, 10Ks, Who's Who in Industry and Finance, and the Wall Street Journal. Although it would have been preferable theoretically to distinguish between "voluntary" departures and "dismissals," the Wall Street Journal and other published sources for CEO departure are either incomplete or unreliable for a large number of firms in the sample, particularly for smaller companies. The proportion of outside directors was estimated as the number of outside board members divided by the total number of directors. The number and affiliation of board members was determined from Standard and Poor's Directory of Corporations, Officers, and Directors, proxy statements, 10Ks, and Who's Who in Industry and Finance.
I included control variables for firm size, characteristics of the CEO (tenure, age, year of hire), and separate board chairman. All control variables, except employment data, were recorded at the beginning of the year. Data on firm size were obtained from COMPUSTAT. All other control variables were obtained from Standard and Poor's Directory of Corporations, Officers, and Directors, supplemented by data from proxy statements, 10Ks, annual reports, and Who's Who in Industry and Finance. Tenure measured the number of years the incumbent had served as CEO. Tenure was used as the duration measure in all models estimated. Consequently, the constant in the b vector served as the coefficient for CEO tenure in all the models. Data on CEOs appointed prior to 1960 are subject to left-truncation, as data sources reveal when the tenure spells began, but the study does not include observations prior to 1960. For firms in the sample, all CEO-years were included, beginning in 1960, but the tenure prior to 1960 was excluded. To address the problem of left-truncation, data on prior CEO tenure were recorded for all incumbents in 1960 or for the first incumbent in the sample for each company, and only the remainder of the CEO's tenure was included in the sample. According to Tuma and Hannan (1984), this procedure leads to consistent estimates. The results of this procedure are conditional on CEOs having survived to 1960.
Size was measured as the logarithm of the number of employees (in thousands). The expectation is that larger firms will have a greater degree of insider CEO succession. Age60 takes the value of zero if the CEO is 60 years old or under, and age minus 60, otherwise, where age is the current age of the CEO. The Age60 variable assumes that up to the sixtieth year of age, age has no effect on CEO succession or selection but that it has increasing effects for each year afterward. Age60 is expected to have a positive effect on the rate of insider CEO succession. Chairman is a dummy variable that takes the value of 1 when the chairman of the board of directors differs from the CEO, and 0, otherwise. First year is a dummy variable for the first year of the tenure of the incumbent CEO.
Table 1 presents the sample means, standard deviations, and the Pearson correlation coefficients for the variables used in the analysis. The unit of observation is the company-year.
I tested the hypotheses by specifying continuous-time, event history analysis of the competing risks of insider versus outsider CEO succession. This competing-risk model estimates the hazard rates of two mutually exclusive events for each incumbent CEO, insider succession and outsider succession. Although succession is a repeatable event from the perspective of the firm, it is non-repeatable from the perspective of the incumbent. Consequently, in the current methodology, the time clock for executive succession refers to the tenure of the incumbent CEO, and the clock is reset to zero when a succession occurs.
This methodology allows for estimating the separate mechanisms that yield insider and outsider succession. It assumes that at any point in time the incumbent CEO is subject to two competing hazards, insider succession and outsider succession, with their own respective contributing factors. The competing-risk models are consistent with the underlying assumption of this paper that the enabling function of the rules of CEO succession affect both the likelihood of a succession [TABULAR DATA FOR TABLE 1 OMITTED] event and the selection of an insider versus an outsider as a CEO. A chi-square contrast of the competing-risk model with the pooled model of combined insider and outsider CEO succession supports the contention that rules of CEO succession affect the overall rate of succession as well as the choice of insider versus outsider and is available from the author. The use of event history analysis of CEO succession is preferable to the more common method of sampling only for the year of turnover, which is subject to sample selection bias and implicitly assumes equilibrium in the CEO succession process (Tuma and Hannan, 1984). The models were estimated using maximum likelihood with the software package TDA (Rohwer, 1991). TDA allows estimation of models with time-varying covariates and takes right-censoring into account by using the information provided by the cumulative survival time of censored spells (Tuma and Hannan, 1984).
The use of event history analysis requires that specification of the functional forms for the transition rates be estimated. I estimated alternative functional forms that allow for either monotonic or nonmonotonic duration dependence, including the exponential, the piecewise exponential, Weibull, Gompertz, and the log normal. The piecewise exponential model provided a better fit than the exponential, consistent with a nonmonotonic duration dependence. Space limitations preclude me from presenting the alternative models estimated, but the comparison of the various models is available from the author. The results of the hypothesis testing are robust across the various specifications of functional form for the duration termu Following Ocasio (1994), I selected the log normal specification, which minimized the log likelihood-ratio.
The hazard rate, r(t) of CEO succession is defined under the two competing risks of insider and outsider succession r[(t).sub.1] and r[(t).sub.2], respectively, as follows:
[Mathematical Expression Omitted], where [Mathematical Expression Omitted],
[Mathematical Expression Omitted], where [Mathematical Expression Omitted],
[Mathematical Expression Omitted], [Mathematical Expression Omitted], [Phi] ([z.sub.t]) is the density function for the standard normal distribution, and [Phi] ([z.sub.t]) is the cumulative density function for the standard normal distribution. The X variables are the time-independent covariates, and the Z variables are the time-dependent covariates. I included year of hire in the b vector to account for the possibility that its effect on the rate of succession varies according to the length of tenure of the CEO.
[TABULAR DATA FOR TABLE 2 OMITTED]
Table 2 summarizes the hypothesized effects, as discussed above. Table 3 presents the results of the competing risks of insider versus outsider CEO succession for incumbent CEOs and provides tests for hypotheses 1-2. Model 1 shows the baseline model for comparison, including the control variables and the main effects of performance, firm age, and year of hire. Models 2-5 present four alternative specifications of the effects of past precedents on CEO succession. All four models include the effects of formal ILMs. Model 2 compares the case of founders, the absence of experience with succession, with any succession experience as the omitted category. In model 3, the case of past insiders is compared with all other categories; in model 4, the case of past outsiders is compared with all other categories. In model 5, categorical variables are included to account for the cases of past insiders, past outsiders, both insiders and outsiders, and founders. In model 5, past outsiders is the omitted category for the rate of insider succession, and past insiders is omitted for the case of outsider succession.(6) The results of models 2, 3, 4, and 5 with chi-square contrasts of 23.06, 22.78, 21.14, and 34.88 with the baseline model 1 (with 4, 4, 4, and 8 degrees of freedom, respectively), are all statistically significant at the .001 level. These results provide initial strong support for the influence of formal and informal rules of CEO succession.
The findings strongly support H1, that formal rules of insider CEO succession, in the form of ILMs for CEOs, increase the rate of insider succession. Models 2-5 show that formal ILMs significantly increase the rate of subsequent insider succession, and these effects are significant at the .001 level. The existence of formal rules of insider succession increases the rate of insider CEO succession, consistent with the enabling function of formal rules.
The findings support H2b and H2c and provide partial support for H2a. In model 2, the effect of founders supports H2a for the case of insider CEO succession; the effect on outsider succession, while in the expected direction, is not statistically significant. Founders have a lower rate of insider CEO succession than all other CEOs, significant at the .01 level. H2b and H2c are supported in model 3, with precedents of insider succession increasing subsequent insiders, and past outsiders increasing outsider succession. Informal rules of both insider and outsider succession affect board decisions on CEO succession. Model 5 compares the effects of all categories of precedents of succession, with past outsider as the omitted category for insider succession and past insider for outsider succession. The results of model 5 also support H2b and H2c. The effects of past insiders on subsequent insider succession are statistically different from the omitted category of past outsiders, consistent with H2b. The effect of past outsider on subsequent outsider succession is statistically different from the omitted category of past insiders, consistent with H2b. Similar effects, not reported here but available from the author, are also obtained if model 5 is estimated omitting all cases of founders. Neither the category of founders nor both insiders and outsiders is statistically different from the omitted categories of past outsider and past insider for either subsequent insider succession or outsider succession. Founders can be shown to be statistically significantly different, however, from past insiders on the rate of insider succession and statistically significantly different from past outsiders on the rate of outsider succession.
Figures 1 and 2 present graphically the effects of the formal ILMs for CEOs and past succession experience on the competing [TABULAR DATA FOR TABLE 3 OMITTED] risks of insider versus outsider CEO succession, respectively. The hazard rates are estimated from model 5, with all the control variables estimated at sample means. The graphical analysis shows the size of the effects of the maximum likelihood estimates. Given the log-normal functional form, hazard rates of both insider and outsider succession are shown as a nonmonotonic function of the duration variable, CEO tenure. The pattern of findings shows that both formal and informal rules of CEO succession have large, substantively significant effects on board decisions. Figure 1 shows that founders have the lowest rate of insider CEO succession, with a hazard rate of .02 or less for CEOs with tenure of less than 15 years. Both formal ILMs in the case of founders and past insider experience with no ILM increase the hazard rate of insider succession to between .04 and .06 for CEOs of 10 years tenure or more. The largest effects are found when boards are subject to both formal ILMs and informal rules of insider succession. In this instance, the rate of succession exceeds .10 for CEOs with tenure between 5 and 30 years, approximately five times higher than for founders. Figure 2 shows the effects of formal and informal succession rules on outsider succession. A comparison of rates of .01 for outsider succession for founders and approximately .04 for past outsider succession for CEOs with tenure of more than 5 years shows that the formal and informal rules of succession significantly transform the CEO succession process and substantially increase the likelihood that the board will adopt either insider or outsider successions, in accordance with the prevailing CEO succession rules.
Moderators of Reliance on Rules of Succession
Table 4 presents the effects of reliance on CEO succession rules under varying organizational contingencies, and table 5 presents results of models testing alternative explanations. As in model 5, for all models except model 10, past outsider was the omitted category for the rate of insider succession, and past insiders was omitted for the rate of outsider succession. Model 10, concerned with showing the interaction effects of founders and formal ILMs estimates the main effect for founders and has all other categories as omitted, as in model 2.
Model 6 shows the moderating effects of firm age on the reliance on rules of CEO succession. The results support an [TABULAR DATA FOR TABLE 4 OMITTED] increased reliance on rules of CEO succession as firm age increases for the case of past insiders (H3a), but no support for the other cases. The chi-square contrast between models 6 and 5 is 7.96 (d.f. = 4, p [less than] .10). The only interaction term that is statistically significant is past insider. The moderator effect of firm age on the reliance on rules of insider succession [TABULAR DATA FOR TABLE 5 OMITTED] indicates that as the firm grows older, its commitment to insider succession increases, which is consistent with the political effects of the rules of insider succession. Reliance on the formal rules of CEO succession is also invariant to firm age.
Model 7 estimates the moderating effects of the prevalence of internal labor markets for CEOs on the rates of insider succession. H4 is not supported, and the maximum likelihood estimate has the opposite sign from the prediction. The chi-square contrast between model 7 and model 5 is also not significant. This suggests that the effects of formal rules of CEO succession do not increase with the degree to which ILMs have been prevalent in the past. Instead, it suggests that formalization, rather than past history, explains reliance on formal ILMs.
Model 8 estimates the moderating effects of year of hire on the reliance on formal and informal rules of CEO succession. The results provide support for H5b, that reliance on past precedents of insider succession will decrease over time, as the relative dominance of insider succession decreases in the U.S. industrial sector. The moderating effects of year of hire on formal ILMs or past precedents of outsider succession, although in the expected direction (H5a and H5c), are not statistically significant, but the main effects are also no longer significant. The chi-square contrast between model 8 and model 5 is not statistically significant. These results, while not conclusive, suggest that the effects of organizational-level rules are relatively invariant to the historical changes in the broader institutional sector.
Model 9 (in table 5) shows the effects of performance on persistence of the formal and informal rules of CEO succession and on insider and outsider succession. The chi-square contrast between models 9 and 5 (in table 3) is not significant. No support is found for H6a, H6b, or H6c, based on bounded rationality, as the interaction effects of ROA and formal ILM, past insiders and past outsiders are not statistically significant. The findings show that formal and informal rules are likely to persist under both positive and negative performance outcomes. Formal ILMs are invariant to performance effects, with statistically insignificant interaction effects with both insider and outsider succession. The effects of past precedents of insiders and outsiders also persist independent of performance, with significant main effects of past insider on subsequent insider succession and of past outsiders on subsequent outsider succession.
The effects of performance on informal CEO succession rules show little evidence that performance triggers a failure-induced change in the application of dominant rules and are more consistent with a view that boards rely on formal and informal rules for their CEO succession decisions. The negative interaction effect of past insider and ROA on insider succession, although not significant, is inconsistent with adaptive change in the rules of insider succession. The interaction effect of past outsider and ROA on outsider succession is positive but not significant. Furthermore, this effect when added to the main effect yields a negative effect on outsider succession, implying persistence with outsider succession rules under economic adversity. Contrary to the theory of bounded rationality, reliance on rules of outsider succession does not change under poor economic performance. The effects show that for firms with precedents of outsider succession, poor performance is more likely to lead to continued reliance on outsider succession.
Model 10 presents the interaction effects of founders and tenure, as captured in the b vector of the model. The chi-square contrast with model 2 (in table 3) is 14.86 with 2 degrees of freedom, which is statistically significant at the .001 level. But the effects do not support H7, on the moderating effects of founder power on insider and outsider succession. The interaction effect of founders and tenure on succession, while negative as predicted by H7, is not significant. When including the effects of founder tenure, the main effects of founders remains statistically significant, which indicates that this result is shaped by the absence of precedents of insider succession and is not an artifact of founder power. Furthermore, the effect of founder tenure on outsider succession is strongly positive and significant rather than negative, as predicted by H7. This result was not expected and is not easily explained. One possibility is that in the absence of rules of succession, whether formal or informal, firms led by founder CEOs are increasingly unlikely to resolve the problem of succession through inside candidates, as no insider is perceived as comparable to the founder. Therefore, as founder tenure increases, firms increasingly look to outsiders as replacements. The results suggest that the social and economic power of founders cannot, by itself, provide an alternative explanation for H2a. Instead, the findings imply that lack of rules of succession serves as a source of power for founders and leads to decreased rates of insider succession. The results also suggest that as founder tenure increases, outsider succession is increasingly likely.
Model 11 distinguishes the effects of formal and informal rules of succession on CEOs facing retirement age and on younger CEOs. The chi-square contrast between model 11 and model 3 (in table 3) is 18.09 with 6 degrees of freedom, significant at the .01 level.(7) In this model, the main effects of formal ILM and past outsider hold for younger CEOs, significant at the .01 and .05 levels, respectively. The effect of past insider is significant at the .10 level. The effects for CEOs facing retirement are the sum of the main effects and the corresponding interaction effects with late departure.
While the results do not support H8a, H8b, or H8c, they suggest that the effects of formal ILMs on insider succession may be greater for CEOs facing retirements than for younger CEOs. The interaction effect of formal ILMs and retirement age is positive, but has a p value less than .10. This effect coupled with a lower significance level for the main effects of past insiders and a positive albeit not significant effect for the interaction between past insiders and retirement age suggests partial evidence for a weaker version of H8a and H8b. While formal ILMs and past precedents of insider succession increase the rate of insider succession for incumbent CEOs prior to retirement age, these effects (particularly that for formal ILMs) may be greater when CEOs are facing retirement. The results are inconsistent with H8c, however, for the interaction effect of past outsiders and retirement age, albeit not significant, is negative, contrary to H8c.
Model 12 presents the interaction effects of proportion of outside directors with formal ILMs. The interaction effects with past insider and past outsider, although not hypothesized, are also shown. The chi-square contrast with model 5 is not statistically significant; neither are any of the interaction effects. There is little evidence that the reliance on for-real rules of succession derives from either the rational, adaptive strategy of outside directors or their social ties with prospective candidates.
The effects of the control variables are unremarkable and consistent with past research. Tenure, measured as the constant in the b vector of the competing-risks model, shows a nonmonotonic relationship between CEO tenure and the rate of succession. This effect can be best examined through figures 1 and 2. The rate of succession increases during the early tenure of a CEO, with decreasing rates afterwards. Performance (ROA) had a negative effect on both insider and outsider succession, albeit only statistically significant in the latter case. Year of hire was positive and significant for both insider and outsider succession, although larger for the latter, suggesting both a general increase in CEO succession and increasing reliance over more recent CEO cohorts on the selection of outsiders as successors. Chairman was positive and statistically significant for both insider and outsider succession and of similar magnitude. This suggests, consistent with the findings of Cannella and Lubatkin (1993), that while having a separate chairman increases CEO succession, it has a limited effect on the succession of insiders versus outsiders. Large firm size increased insider succession and decreased outsider succession, but the effects on insider succession were not significant across all model specifications. In particular, including past insiders in the model made the size effect not significant. The effects of firm age were mediated by past succession experience, with no significant main effects.
DISCUSSION AND CONCLUSIONS
This paper both applies and extends institutional theories of rule-based action and decision making and quantitative empirical research on their effects to a new area of application, the study of corporate governance and CEO succession. Empirically, the paper extends this line of research by explicitly testing how reliance on rules is moderated by economic performance and organizational contingencies (e.g., early CEO departures) and by distinguishing between formal and informal rules. Theoretically, it extends previous institutional approaches by focusing on the interplay between cognitive and political factors by which rules both enable and constrain decisions by the organization and that lead to their institutionalization. This approach combines neoinstitutionalist perspectives that focus on cognitive mechanisms (DiMaggio and Powell, 1991), with more traditional sociological perspectives that highlight political factors (Gouldner, 1954; Selznick, 1957).
The results provide strong empirical support for directors' reliance on formal and informal rules of CEO succession. Both the adoption of internal labor markets for new CEOs and the past reliance on precedents of insider and outsider succession had large, statistically significant effects on the competing risks of insider versus outsider CEO succession. The formal rules of insider CEO succession strongly increased insider succession, and these effects were relatively invariant to the level of economic performance, firm age, ownership structure, or whether succession occurred under retirement or early CEO departures. The formal internal labor markets for CEOs imply both political and normative commitments to insider succession, and boards were much more likely than not to honor these commitments, independent of the circumstances facing the firm or the CEO succession process.
The findings highlight the importance of historical precedents in shaping subsequent CEO succession. The absence of such precedents, for founders, resulted in lower rates of insider succession. Reliance on past precedents leads to the routinization of the succession process and is closely tied to the routinization of charisma and the depersonalization of executive power. This process embodies and legitimizes the authority of the corporation in its rules and procedures rather than in power and control by specific individuals (Weber, 1978). But unlike the pure form of Weber-Jan bureaucracy, boards rely on historical precedents as well as formal procedures to define appropriate behavior and to guide their decisions.
The results support the view that rules serve both to enable and constrain actions and decisions by boards of directors. The enabling function of rules is shown by the increased rate of succession when formal and informal rules and precedents are in place. Formal and informal rules of insider succession facilitate subsequent insider succession. Precedents of outsider succession permit greater reliance on outsider succession when the firm is faced with poor economic performance or under contingencies that lead to early CEO departure. The constraining function of the rules is shown by the reliance on CEO selection rules under poor economic performance, increased insider ownership, and increased organizational age.
The results provide little evidence that the reliance on rules can be better explained by alternative explanations, including bounded rationality, founder power, or the structure of the board of directors. The results do not support the early March-Simon-Cyert formulation of failure-induced change in organizational rules and routines but are more consistent with the view that responses to economic adversity rely on dominant rules and logics of action (Staw, Sandelands, and Dutton, 1981; Ocasio, 1995). Formal and informal rules of CEO succession serve to empower board members to respond to adversity by employing the existing rules. Organizations change in response to economic adversity, but in ways that are consistent with dominant patterns and rules of behavior. These results support Milliken and Lant's (1991) view that firms will persist with their strategies and routines, independently of whether recent economic performance is good or whether it is poor.
The moderating effects of organizational age on reliance on formal and informal rules contributes to our understanding of how rules become institutionalized. Formalization was found to be relatively invariant to the effects of firm age. The different effects for dominant rules of insider versus outsider CEO succession suggest that internal political support for an activity may be a necessary condition to guarantee its institutionalization as firms grow older (Selznick, 1957). If a rule is at odds with the interests of internal groups, as in rules of outsider succession, its effects are unlikely to increase with firm age. This suggests that the institutionalization of action requires political support as well as historical precedent and that cognitive factors, while important in explaining the institutionalization of action, are not sufficient.
The pattern of results suggests that both cognitive and political factors are involved in reliance on rules of CEO succession. Neither political nor cognitive factors can explain, by themselves, institutionalized action and decision making in corporate governance. Reliance on rules of succession cannot be fully explained by managerialist accounts that focus on insiders trying to maintain their political position and control of the corporation. First, the evidence of reliance on precedents of outsider succession is not easily reconciled with a managerialist interpretation. Second, there is evidence both for the existence of a generalized rule of outsider succession under adversity and for reliance on rules of outsider succession for CEOs departing prior to retirement age. Neither effect is consistent with a purely political explanation of the rules of CEO succession. Furthermore, I also conducted additional analysis, not reported here, of the mediating effects of the political structure of the board, as measured by a separate chairman, the proportion of outside directors, and board size, on the reliance on rules. None of these measures of managerial control over the board of directors had a noticeable or statistically significant effect on the utilization of either formal or informal rules of CEO succession. This strongly suggests that the operation of rules of corporate governance cannot be reduced to an explanation based on the view of boards as "pawns" of management (Lorsch, 1989).
Reliance on rules of CEO succession highlights the commitments and identification of boards of directors with the mission, purpose, and social identification of the corporation (Selznick, 1957). Legal responsibilities of and normative commitments by board members are reflected in board members abiding by the institutional embodiment of purpose in the firm's rules and routines. At the same time, this rule-bound character of board decision making does not imply that rules are always consistently applied, nor that the process is purely apolitical and conflict free. The effects of rules in the transition-rate models strongly support the view that rules matter, and matter greatly in board decisions, but they also show that their application is far from inevitable. The different moderating effects of age and early CEO departures for insider and outsider succession rules imply that historical precedents serve to establish the rules of the game in corporate politics, both as vehicles for and constraints on managerial power. Reliance on rules of CEO succession and corporate governance also suggests that political maneuverings in organizations are unlikely to be directly expressed or mobilized within the board of directors. This is consistent with a view of managerial power and political dynamics driven by emergent power struggles rather than fixed political coalitions and where overt efforts at social influence, managerial entrenchment, and the institutionalization of executive power are not likely to be effective (Ocasio, 1994). Political struggles and maneuverings occur not as part of the formal board process but are driven by inside participants. The institutionalized action in corporate governance drives CEO succession and begins anew in each corporation with the transition from the founding CEO, where personal power is of greater importance, to the establishment of routines of succession, where formal and informal rules serve to guide action and empower and direct board decisions.
I thank Mauro Guillen, Ranjay Gulati, Rebecca Henderson, Paul Hirsch, Jim March, Mark Mizruchi, Jeff Pfeffer, Brian Uzzi, Marc Ventresca, Ed Zajac, Nick Ziegler, seminar participants at Cornell, Illinois, MIT, and Northwestern, and three anonymous ASQ reviewers for comments and suggestions on earlier versions of this paper.
1 It differs, however, from the earlier Carnegie school tradition by focusing on rule-based action rather than on analysis-based action and boundedly rational choice. In the introduction to the second edition of their classic treatise, March and Simon (1993: 7-13) argued that their earlier formulation overstated the importance of choice and understated the importance of rules and rule-based action.
2 DiMaggio and Powell's (1991) cognitive perspective on institutional theory is presented as descriptive of much research to date rather than as desiderata. The importance of interest and agency in institutional perspectives has been highlighted elsewhere by both authors (DiMaggio, 1988; Powell, 1991).
3 Mark Mizruchi, following (implicitly) a logic of consequences, suggested that this hypothesis could be reversed, that boards may decide they want an insider, then they designate inside officers as potential successors, creating formal positions for them. I believe that the alternative interpretation in H1, which follows a logic of appropriateness, is more plausible, however, given that the average ILM has been in place for six years at the time of CEO succession. Otherwise, boards' decisions on insiders versus outsiders would take place without accounting for over half a decade of executive actions, decisions, and corporate outcomes. The logic of consequence argument, could, of course, be maintained under an interpretation of credible commitment. At heart, the differences between the two interpretations is a metatheoretical one, whether rules are rational responses to interests and outcomes, or path-dependent responses to past events, guided by an alternative logic to rationality, the logic of appropriateness (March and Olsen, 1989).
A closely related point is whether formal ILMs for CEOs are but a justification for decisions made on other grounds, but I would argue that this justification only works when formal ILMs are followed by insider succession, as predicted by the logic of appropriateness.
4 This hypothesis can also be derived from explanations based on the founder's power, although the absence of precedents of CEO succession is itself a source of founder power.
5 An anonymous reviewer suggested that outside directors might rely on their recent experiences in making CEO appointments, but data limitations, particularly for the smaller firms, precluded testing this hypothesis. In addition, while this paper has focused on firm-level rules of succession, rules of succession may also be diffused through interlocking boards of directors.
6 Alternatively, founders can be used as the omitted category, with similar results But as indicated by an anonymous reviewer, the statistical significance of the coefficients of past insider or past outsider could then reflect the lower rates of succession for founders and not the relative effects of informal rules of insider versus outsider succession. To verify that the effects of past insider and past outsider are not an artifact of lower rates for founders. I also estimated model 5 excluding all the cases of founders from the analysis. The results are substantively equivalent To those presented here.
7 Most of this effect is due, however, to the main effect of the retirement age variable rather than to the interaction effects. The chi-square contrast between model 11 and an alternative model, not reported here, that includes the retirement age variable but not the interaction effects is 4.60 with 4 degrees of freedom, which is not statistically significant.
Aldrich, Howard E., and Ellen Auster 1986 "Even dwarfs started small: Liabilities of age and size and their strategic implications." In Barry M. Staw and L. L. Cummings (eds.), Research in Organizational Behavior, 8: 156-198. Greenwich, CT: JAI Press.
Amburgey, Terry L., and Anne S. Miner 1992 "Strategic momentum: The effects of repetitive, positional, and contextual momentum on merger activity." Strategic Management Journal, 13: 335-348.
Berle, Adolf, Jr., and Gardiner C. Means 1967 The Modern Corporation and Private Property. (First published in 1932.) New York: Harcourt, Brace, and World.
Bjorkman, Ingmar 1989 Foreign Direct Investments: An Empirical Analysis of Decision Making in Seven Finnish Firms. Helsingsfors: Swedish School of Economics and Business Administration.
Boeker, Warren, and Jerry Goodstein 1993 "Performance and successor choice: The moderating effects of governance and ownership." Academy of Management Journal, 36: 172-186.
Cannella, Albert A., and Michael Lubatkin 1993 "Succession as a sociopolitical process: Internal impediments to outsider selection." Academy of Management Journal, 37: 763-793.
Carroll, Glenn R. 1984 "Dynamics of publisher succession in newspaper organizations." Administrative Science Quarterly, 29:93-113.
Chandler, Alfred D. 1977 The Visible Hand: The Managerial Revolution in American Business. Cambridge, MA: Harvard University Press.
Clark, Robert C. 1991 "Agency costs versus fiduciary duties." In John W. Pratt and Richard J. Zeckhauser (eds.), Principals and Agents: The Structure of Business: 55-79. Boston: Harvard Business School Press.
Cohen, Michael D., James G. March, and Johan P. Olsen 1972 "A garbage can model of organizational choice." Administrative Science Quarterly, 17: 1-25.
Cyert, Richard M., and James G. March 1963 A Behavioral Theory of the Firm. Englewood Cliffs, NJ: Prentice-Hall.
DiMaggio, Paul D. 1988 "Interest and agency in institutional theory." In Lynn G. Zucker (ed.), Institutional Patterns and Organizations: 3-22. Cambridge, MA: Ballinger.
DiMaggio, Paul D., end Walter W. Powell 1983 "The iron cage revisited: Institutional isomorphism and collective rationality in organizational fields." American Sociological Review, 48: 147-160.
1991 "Introduction." In W. W. Powell and P. J. DiMaggio (eds.), The New Institutionalism in Organizational Analysis: 1-38. Chicago: University of Chicago Press.
Doeringer, Peter B., and Michael J. Piore 1971 Internal Labor Markets and Manpower Analysis. Lexington, MA: Heath Lexington Books.
Fama, Eugene L., and Michael C. Jensen 1983 "Separation of ownership and control." Journal of Law and Economics, 26: 301-325.
Fligstein, Neil 1990 The Transformation of Corporate Control. Cambridge, MA: Harvard University Press.
Fredrickson, James W., Donald C. Hambrick, and Sara Baumrin 1988 "A model of CEO dismissal." Academy of Management Review, 13: 255-271.
Gibbons, Robert, and Kevin J. Murphy 1990 "Relative performance evaluation for chief executive officers." Industrial and Labor Relations Review, 45: 30S-51S.
Giddens, Anthony 1984 The Constitution of Society. Berkeley: University of California Press.
Gouldner, Alvin W. 1954 Patterns of Industrial Bureaucracy. New York: Free Press.
Grusky, Oscar 1961 "Corporate size, bureaucratization, and managerial succession." American Journal of Sociology, 47: 261-269.
Hannan, Michael T., and John Freeman 1984 "Structural inertia and organizational change." American Sociological Review, 49: 149-164.
1989 Organizational Ecology. Cambridge, MA: Harvard University Press.
Helmich, Donald L., and Warren B. Brown 1972 "Successor type and organizational change in the corporate enterprise." Administrative Science Quarterly, 17: 371-381.
Hirsch, Paul 1982 "Network data versus personal accounts: The normative culture of interlocking directorates." Unpublished manuscript, Graduate School of Business, University of Chicago.
Jackall, Robert 1988 Moral Mazes: The World of Corporate Managers. New York: Oxford University Press.
Jensen, Michael C. 1993 "The modern industrial revolution, exit, and the failure of internal control systems." Journal of Finance, 48: 831-880.
Jensen, Michael C., and William H. Meckling 1976 "Theory of the firm: Managerial behavior, agency costs and ownership structure." Journal of Financial Economics, 3: 305-360.
Kelly, Dawn, and Terry L. Amburgey 1991 "Organizational inertia and momentum: A dynamic model of strategic change." Academy of Management Journal, 34: 591-612.
Lorsch, Jay W. 1989 Pawns or Potentates. Boston: Harvard Business School Press.
Mace, Myron L. 1971 Directors: Myth and Reality. Boston: Harvard Graduate School of Business Administration.
March, James G. 1994 A Primer on Decision Making: How Decisions Happen. New York: Free Press.
March, James G., and Johan P. Olsen 1976 Ambiguity and Choice in Organizations. Bergen, Norway: Universitesforlaget.
1989 Rediscovering Institutions: The Organizational Basis of Politics. New York: Free Press.
March, James G., and Herbert A. Simon 1958 Organizations. New York: Wiley.
1993 Organizations, 2d ed. Cambridge, MA: Blackwell Business.
McEachern, William A. 1975 Managerial Control and Performance. Lexington, MA: Heath.
Meyer, John W., and Brian Rowan 1977 "Institutionalized organizations: Formal structure as myth and ceremony." American Journal of Sociology, 83: 340-363.
Milliken, Frances J., and Theresa K. Lant 1991 "The impact of an organization's recent performance history on strategic persistence and change." In Paul Shrivastava, Anne Huff, and Jane Dutton (eds.), Advances in Strategic Management, 7 129-156. Greenwich, CT: JAI Press.
Nelson, Richard R., and Sidney G. Winter 1982 An Evolutionary Theory of Economic Change. New York: McGraw-Hill.
Ocasio, William 1994 "Political dynamics and the circulation of power: CEO succession in U.S. industrial corporations, 1960-1990." Administrative Science Quarterly, 39: 285-312.
1995 "The enactment of economic adversity: A reconciliation of theories of failure-induced change and threat-rigidity." In L. L. Cummings and Barry M. Staw (eds.), Research in Organizational Behavior, 17: 287331. Greenwich, CT: JAI Press.
Pfeffer, Jeffrey, and Gerald R. Salancik 1978 The External Control of Organizations. New York: Harper and Row.
Powell, Walter W. 1991 "Expanding the scope of institutional analysis." In W. W. Powell and P. J. DiMaggio (eds.), The New Institutionalism in Organizational Analysis: 183-203. Chicago: University of Chicago Press.
Puffer, Sheila M., and Joseph B. Weintrop 1991 "Corporate performance and CEO turnover: The role of performance expectations." Administrative Science Quarterly, 36: 1-19.
Rohwer, Gotz 1991 "Analysis of transition data: A practical introduction with TDA." Unpublished manuscript, European University Institute.
Scott, W. Richard 1995 Institutions and Organizations. Thousand Oaks, CA: Sage.
Selznick, Philip 1957 Leadership in Administration. Evanston, IL: Row, Peterson.
Simon, Herbert A. 1957 Models of Man: Social and Rational. New York: Wiley.
1997 Administrative Behavior: A Study of Decision-Making Processes in Administrative Organizations, 4th ed. New York: Free Press.
Staw, Barry M., Lloyd E. Sandelands, and Jane E. Dutton 1981 "Threat-rigidity effects in organizational behavior: A multilevel analysis." Administrative Science Quarterly, 26: 501-524.
Tuma, Nancy B., and Michael T. Hannan 1984 Social Dynamics: Models and Methods. Orlando, FL: Academic Press.
Tushman, Michael L., and Elaine Romanelli 1985 "Organizational evolution: A metamorphosis model of convergence and reorientation." In L. L. Cummings and Barry M. Staw (eds.), Research in Organizational Behavior, 7: 171-222. Greenwich, CT: JAI Press.
Useem, Michael 1984 The Inner Circle. Oxford: Oxford University Press.
1996 Investor Capitalism: How Money Managers Are Changing the Face of Corporate America. New York: Basic Books.
Vancil, Richard M. 1987 Passing the Baton: Managing the Process of CEO Succession. Cambridge, MA: Harvard Business School Press.
Walsh, James P., and James K. Seward 1990 "On the efficiency of internal and external corporate control mechanisms." Academy of Management Review, 15: 421-458.
Weber, Max 1978 Economy and Society. Guenther Roth and Claus Wittich, eds. Berkeley: University of California Press.
Weisbach, Michael S. 1988 "Outside directors and CEO turnovers." Journal of Financial Economics, 20: 431-460.
Westphal, James, and Edward J. Zajac 1995 "Who shall govern?: CEO/board power, demographic similarity, and new director selection." Administrative Science Quarterly, 40: 60-83.
Zajac, Edward J. 1990 "CEO selection, compensation, and firm performance: A theoretical interpretation and empirical analysis." Strategic Management Journal, 11: 217-230.
Zhou, Xueguang 1993 "The dynamics of organizational rules." American Journal of Sociology, 98: 1094-1133.
Zucker, Lynne G. 1977 "The role of institutionalization in cultural persistence." American Sociological Review, 42: 726-743.
|Printer friendly Cite/link Email Feedback|
|Publication:||Administrative Science Quarterly|
|Date:||Jun 1, 1999|
|Previous Article:||Psychological safety and learning behavior in work teams.|
|Next Article:||The Structure of Women's Nonprofit Organizations.|