The Role of Chief Executive Officer

Theories of Leadership

The leadership role is believed to be one of the most important roles of a CEO with a reach that spans all other roles. This belief makes a brief description of leadership theories necessary for informing research on CEOs (Goleman, Boyatzis, & McKee, 2002; Mintzberg, 1973; Steiner, Kunin, & Kunin, 1981). CEOs perform in the upper echelon of the organizations they lead. Upper echelon theory helps to explain the influence of top leaders on organizational development, suggesting CEOs and other top leaders are able to reflect their thoughts and values into an organization. This ability is based on how much discretion these leaders have to act independently, that is, CEOs who act independently are more able to influence the organization (Boal & Hooijberg, 2001; Hambrick & Mason, 1984; Hiller, Resick, Weingarden, & Whitman, 2009). Upper echelon theory evolved into strategic leadership theory, which suggests organizations are truly reflections of their top leaders. Taking these theories one step further, positive agency theory purports leaders act in their own best interests, as opposed to the best interests of the organization (Boal & Hooijberg, 2001). To increase alignment between the goals of leaders and those of stakeholders, boards of directors use incentive systems tied to organizational performance to influence top leadership (Boal & Hooijberg, 2001). Leadership theory in general describes leaders at all levels of an organization (not necessarily CEOs); however, strategic leadership theory refers only to top organizational leaders (Vera & Crossan, 2004).

Strategic Leadership Theory

Unique to strategic leadership theory are the concepts of adaptive capacity, or the ability to change (Cummings & Worley, 2009), and absorptive capacity, or the ability to learn. Absorptive capacity requires that the individual constantly experiment, tolerate small failures, and engage in double-loop learning (Argyris & Schon, 1978; Boal & Hooijberg, 2001). Handling the rapid change and complexity of today’s business environment suggests that leaders need to become ambidextrous or to develop “the capacity to simultaneously implement diverse courses of action: incremental and discontinuous innovation, exploration and exploitation, flexibility and control, and feed-forward and feedback learning” (Vera & Crossan, 2004, p. 227). Bodwell and Chermack (2010) suggested, “Ambidexterity eliminates the need for organizations [or CEOs] to choose between opposites or to focus on ‘trade-offs’” (p. 197). From strategic leadership theory transformational leadership evolved, including the concepts of complexity and the ability to handle juxtaposing positions (Bass, 1985, 1998). The transformational leadership style is described in the following section.

Transformational Leadership

The transformational leader stands in stark contrast to the transactional leader who focuses on internal processes and managing others to get the job done (Bass, 1990; Vera & Crossan, 2004). Instead, the transformational leader “asks followers to transcend their own self-interests for the good of the group, organization, or society; to consider their longer-term needs to develop themselves, rather than their needs of the moment; and to become more aware of what is really important” (Bass, 1990, p. 53). Transformational leaders are identified as being “charismatic, inspirational, intellectually stimulating, and individually considerate” (Vera & Crossan, 2004, p. 224). Described as selfless or servant leadership, transformational leadership focuses on the development of all members of an organization into leaders (Bass, 1990). Emerging theories resulting from transformational leadership include the complexity theories of behavioral and cognitive complexity introduced by Hart and Quinn (1993) and explored in more detail by Zaccaro (2001). Transformational leadership has a focus on integrity; however, integrity alone will not insure strong organizational performance. Leaders often struggle with the desire to do the right thing while at the same time making sure the organization performs financially. Lynham’s (1998, 2000) theory of responsible leadership for performance addresses this tension.

Theories of Responsible Leadership for Performance

Ambiguity in theories of leadership calls for both firm performance and selflessness (Lynham, 1998). The selfishness of human nature conflicts with the ability to be a servant leader, especially when complicated by the demand for corporate performance from shareholders and boards of directors. It is not difficult to imagine how executives become tempted to force earnings when their personal wealth can be substantially influenced by a strong stock price (Boatright, 2009; Jensen & Murphy, 1990). To resolve the conflict between selfish needs and the needs for performance, Lynham (1998) has suggested a new model of “Responsible Leadership for Performance” (p. 208). The need for a new model stems from the leadership dilemma that remains in many organizations, lack of evidence that leadership development programs sustain real results, increased complexity in business, and increased rate of change in which leaders must act (Antonioni, 2003; Block, 1993; Lynham, 1998; Zaccaro, 2001).

Gone are the days of prescriptive recipes of leadership. Organizations will need leadership that is able to think and act fundamentally differently, able to integrate critical thinking with critical practice. Organizations will need to shift away from the pursuit of control, discrete boundaries and recognizable problems towards greater complexity, global competition, continuous quantum change and collaborative team orientations. These shifts herald an increasing gap between current leadership practices and future leadership needs. The competencies needed of and for leadership will be different. They will need to integrate the apparent paradox of performance demands and the realities of humanity—in other words, of people and performance. (Lynham, 1998, p. 209)

Lynham (1998) also suggested that leadership systems should be developed to answer the problem of succession planning. It is not uncommon for organizations to undergo substantial changes due to the change of one individual at the top. This is disruptive and often very unproductive for the organization (Case, 2009). A recent article about the replacement of HP CEO Mark Hurd suggested that the CEO position should be a “team job” arguing that

a rotational CEO position would put boards in control of the CEO spot, rather than the other way around. Instead of paying huge sums to—and relying exclusively on—a single individual, boards would groom multiple individuals of diverse skills useful for finite durations, with paychecks to match. Investors would benefit knowing the company had several qualified and/or tested CEOs from which to choose. Perhaps, by failing to engage in adequate succession planning and to rein in pay, boards have, inadvertently, pointed us to a better, more cost-effective model at the top. (Bloxham, 2010, p. 3)

The Contribution of Leadership Theories in Informing Research on the Role of CEO

The theories of leadership described in the literature suggest that the leadership role of CEOs is an evolving role. The newer models suggest CEOs are leaders not only within their organizations but also globally, and, as global citizens, the role is even more important (Antonioni, 2003; Matsumura & Shin, 2005). It seems the need for these evolved models of leadership could not have been imagined by Mintzberg during the 1960s or 1970s. Furthermore, he could not have anticipated the magnitude of changes that would occur in the 40 years following his research on five CEOs, yet his study continues to provide a conceptual framework for contemporary research on the role of CEO. His work was an attempt to create a theory about the role of CEO by observing what CEOs did at work. A description of his research and the resulting theory on the role of CEO is presented in the following section.

Mintzberg’s Methodology: Structured Observation

Mintzberg’s intention in applying the structured observation method was the inductive development of a theory about the role of managers (Mintzberg, 1973). Specifically, he observed five male CEOs: (a) a chairman and chief executive from a major consulting firm, (b) a president of a research and development firm engaged in high-technology solutions for industry and the military, (c) a director of a large urban hospital, (d) a president of a manufacturing company producing consumer goods, and (e) a superintendent of a large suburban school district (Mintzberg, 1973).

Each subject was observed for 1 week while at work. The observations resulted in a chronology record, a mail record, and a contact record. The chronology record summarized the amount of time each participant spent on desk work, telephone calls, scheduled meetings, unscheduled meetings, and tours. The mail record summarized the type and amount of mail each subject received and read at work during the week of observation. The contact record summarized the type of contacts: (a) verbal, (b) telephone, (c) scheduled meeting, (d) unscheduled meeting, and (e) who the contact was: subordinates, directors, clients, suppliers, peer/trade organizations, independents, and other. The purpose of the contact was categorized as organizational, scheduling, ceremony, board work, status requests and solicitations, action requests, manager requests, observational tours, receiving information, giving information, review, strategy, and negotiation. The activities from each record were categorized and, from the process of categorization, Mintzberg’s 10 roles were delineated.

The review of literature contributed to this study in two significant ways. The research revealed 31 roles that are the basis for the survey instrument developed for this research study with the purpose of answering Research Question 1. Furthermore, the review of literature and the process of constant comparative analysis resulted in the identification of six major role categories comprising the role of CEO. These six role categories provide a systematic way of synthesizing the research on the 31 roles and create a simplistic system for understanding and thinking about the role of CEO. These role categories are used in the survey instrument to answer Research Question 2.

The six role categories consist of informational, interpersonal, decisional, operational, strategic, and diplomacy roles. They are presented in detail in the following section and were developed using constant comparative analysis. Constant comparative analysis “is an inductive (from specific to broad) data analysis procedure in grounded theory research of generating and connecting categories by comparing incidents in the data to other incidents, incidents to categories, and categories to other categories” (Creswell, 2002, p. 451). This analysis was made by reviewing the detailed descriptions of each role presented in the literature and then grouping the similar descriptions together, forming the categories of roles. The discussion of the six role categories begins with the informational roles.

Sampling Procedures

The sampling procedures for this study were conducted in phases aimed at maximizing the number of study participants. The first phase was composed of a convenience sample of 127 CEOs. The advantage of using a convenience sample is that it is believed CEOs are more likely to respond to surveys when they are conducted by someone they know or trust (Bartholomew & Smith, 2006). The second phase of sampling was based on an accessible population of CEOs from a purchased database. The entire accessible population was contacted. A large sample was necessary due to the expected low response rate.

External Validity

The goal of sampling is to select a sample that is representative of the total population so that inferences can be made from the sample to the population (Gall, Gall, & Borg, 2005; Gliner, Morgan, & Leech, 2009). Having a sample that is representative is more important than overall sample size and is an important aspect in evaluating external validity, specifically population external validity (Gliner et al., 2009). Ecological external validity, a second aspect of external validity, is considered adequate when the study is conducted in a natural setting for the participants (Gliner et al., 2009). Due to the nature of a mailed or e-mailed survey, it is assumed CEOs who choose to participate will do so from their homes or their offices—a natural setting for them as opposed to a laboratory or classroom. For purposes of this study, internal validity will not be evaluated because the research is not designed to infer a causal relationship (Gliner et al., 2009).

Measurement Validity and Reliability

Measurement validity establishes evidence for using a particular instrument in a “particular setting with a particular population for a specific purpose” (Gliner et al., 2009, p. 165). Validity is based on the scores obtained from the instrument. There are three types of evidence supporting measurement validity of an instrument: content validity, criterion-related validity, and construct validity (Gall et al., 2005; Gliner et al., 2009). Measurement validity is more difficult to obtain than measurement reliability. Both reliability and consistency are necessary for measurement validity (Gliner et al., 2009).

Content evidence can be provided by using an instrument that has already been used or by having experts review an instrument to verify its content represents what the instrument is trying to measure. This instrument has not been used previously; however, these roles emerged from research over the past 40 years and, therefore, it is believed there is some level of content evidence for the roles on the survey. The use of expert CEOs to review the instrument also provides content validity. The process of content analysis was used to induce the six categories of roles by analyzing the descriptions of the roles and grouping the descriptions that were most similar into the categories described in the literature.

Construct validity evidence can be provided by response processes in which participants do not answer with responses they believe to be socially desirable. It can also be provided with evidence based on internal structure, which is tested via factor analysis. For purposes of this study, Cronbach’s alpha was used to assess internal consistency reliability of the roles within the major categories: informational, interpersonal, decisional, operational, strategic, and diplomacy. Specifically, factor analysis and Cronbach’s coefficient alpha is used for testing internal consistency reliability by measuring consistency in a multiple-item scale when enough responses are received (Leech, Barrett, & Morgan, 2005). Cronbach’s alpha is usually used “when the researcher has several Likert-type items (ratings from strongly disagree to strongly agree) that are summed or averaged to make a composite score or summated scale” (Gliner et al., 2009).

This instrument was tested using factor analysis. Factor analysis helps determine which items result in consistent responses from survey participants. Consistency in responses suggests certain items are measuring the same construct and fit the categories (Gall et al., 2005). The alpha is then based on the average correlation of each item with every other item and is used because it provides a measure of reliability based on one administration of a questionnaire or survey (Gliner et al., 2009). In this respect, the use of alpha could significantly reduce the number of items on the survey for future use.

Description of Sample

The sample for this study was obtained through professional relationships and two databases of U.S. companies containing CEO e-mail addresses. The accessible population comprised 28,018 possible study participants; however, two e-mails were returned due to invalid e-mail addresses. After receiving IRB approval, CEOs were e-mailed a link to the survey along with the appropriate IRB consent information. A total of 1,768 surveys were started, 1,237 were completed, and 1,202 were considered usable, for a total response rate of 4.29%.

Nonrespondents

Nonrespondents totaled 26,784. This is a substantial, but not unexpected, number of nonresponders. According to Dillman (2007), nonresponse error is one source of error that results from survey research. Other errors include sampling error, coverage error, and measurement error. An increase in any one of these types of error may cause the results of a study to “become increasingly suspect and decreasingly valuable as evidence of the characteristics in other audiences” (Dooley & Lindner, 2003, p. 100). There are several recommended solutions for controlling for nonresponse error after follow-up procedures have been used. These include “ignore nonrespondents; compare respondents to population on characteristics known before the study; compare respondents to nonrespondents on characteristics known before the study; compare early to late respondents; and ‘double-dip,’ or sample, nonrespondents” (Dooley & Lindner, 2003, p. 102).

Lindner, Murphy, and Briers (2001) suggested comparing the last wave of 30 respondents to early respondents as one defensible and “generally accepted procedure(s) for handling nonresponse error as a threat to external validity of research findings” (Dooley & Lindner, 2003, p. 103). In an independent sample t test of the first 30 and last 30 responses received, there were no statistically significant differences on agreement with role categories or time spent in the role categories between the two groups sampled.

Demographics

The sample is composed of 1,048 males (87.5%) and 150 females (12.5% total, and 15 from public companies). The S&P Fortune 500 (Lipman, 2011) contains approximately 13 female CEOs, making this sample seem fairly representative in terms of gender. The majority of respondents range in age from 45 to 64, with 21.3% between 50 and 54, 20.9% between 55 and 59, and 18.7% between 60 and 64. The age distribution was also similar to the age distribution of CEOs in S&P Fortune 500 companies. Average tenure in current job is 12.4 years and the average tenure as a CEO is 11.6 years. Average company size was 962 employees and US$253 million in sales revenues.

Exploratory Factor Analysis (EFA)

The primary research focus was not instrument development; however, as the survey instrument used was designed specifically for this study, it is appropriate to run EFA to examine the reliability and validity of the instrument (Yang, 2005). The purpose of EFA is “to discover the common factors that drive interrelationships among the observable variables” (Yang, 2005, p. 184). EFA is preferable over confirmatory factor analysis (CFA) “in early stages of scale development . . . to explore the underlying factor structure and to determine how measurement items load on factors that have not been clearly revealed” (Yang, 2005, p. 185). Principal components analysis (PCA) was run with Varimax (orthogonal) rotation. Rotation can be described as “mathematical alignment” of variables in which “variables that cluster closely together on some axis are presumably related to each other” (Leong & Austin, 2006, p. 251). Rotation is used “to improve the psychometric properties (i.e. reliability and validity) and substantive meanings of extracted factors” (Yang, 2005, p. 192). Principal axis factor analysis (PAA) was also run with Promax (oblique) rotation and the results were not substantially different. PCA was used for this study as PCA is the most commonly used form of factor analysis (Leong & Austin, 2006). The sample size was adequate for factor analysis. Leong and Austin (2006) recommended giving a survey to at least five times as many people as there are questions on the survey.

PCA revealed seven components comprising 52.4% of the total variance explained. Using a loading of .40, appropriate for a new survey (Yang, 2005), the 31 items sorted into seven components with 27 items retained. Three of the roles, entrepreneur, conflict handler, and consultant, did not load. The fourth excluded role, commander, did not fit logically with the diplomacy roles and the decision was made to exclude the role.

Research Question 1

Research Question 1 asked, “What is the role of CEO in the United States?” Almost 99% of participants strongly agreed or agreed that leader (Mintzberg, 1973) is a role of the CEO. Also considered the most important role of a CEO (Edersheim, 2007; Lafley, 2009; Mintzberg, 1973; Quarterman et al., 2005), this role was described on the survey as “I lead and motivate my subordinates.” The leader role was part of the interpersonal role categories, which included motivator (96% of respondents strongly agreed or agreed with this description—“I create and set a sense of excitement and vitality in the organization, challenging people to gain new competencies and achieve higher levels of performance”; Hart & Quinn, 1993) and director (90.3% strongly agreed or agreed with this description—“I make sure the right people are in the right place at the right time doing the right things”; Gulick, 1937). The leadership roles resonated with participants because leader may be the image most commonly associated with CEO. In the literature, leader was designated as one of the most important roles of a CEO, touching all other roles within an organization (Goleman et al., 2002; Mintzberg, 1973; Steiner et al., 1981). The data obtained in this study support this conclusion. Some participants criticized the survey for the use of “I” in describing the CEO roles, indicating that the role was less focused on the CEO as an individual and more focused on the CEO’s ability to lead and motivate the teams making up their organizations.

The decisional roles also received a large percentage of strongly agreed or agreed responses from the participants. These roles include disturbance handler (96.7% strongly agreed or agreed with this role description—“I take charge when my organization is threatened”; Mintzberg, 1973), taskmaster (96.6% strongly agreed or agreed—“I have a strong focus on results or getting the job done”; Hart & Quinn, 1993), resource allocator (91.5% strongly agreed or agreed—“I decide where my organization will expend efforts and resources”; Mintzberg, 1973), and staffer (87.7% strongly agreed or agreed—“I make sure the right people are hired for the right positions”; Gulick, 1937). Participant alignment with the decisional roles seems natural based on the CEO’s unique authority to guide the direction of the organization either through independent decision making or through the process of influencing consensual decision making. CEOs are perceived as having the final say in guiding the organization (Breene et al., 2007).

The strategic roles also received strong ratings as CEO roles. Vision setter (Hart & Quinn, 1993) received the most strongly agreed/agreed responses (93.5% of participants) to the description, “I create a sense of identity and mission for my organization.” Also included are creator/maintainer of culture (91.6% strongly agreed/agreed—“I establish and ensure the organization’s culture is consistent with its strategic focus and plan”; Sashkin & Fulmer, 1988), coordinator (90.8% strongly agreed/agreed—“I make sure all efforts are coordinated toward the goals and strategic plan of the organization”; Fayol, 1916; Gulick, 1937), innovator (87.9% strongly agreed/agreed—“I guide the organization into new cycles of innovation”; Galambos, 1995), transformer (83.4% strongly agreed/agreed—“I transform the organization as markets and the external environment change”; Galambos, 1995), strategist (83.1% strongly agreed/agreed—“I craft the organization’s strategy”; Stata, 1988), and planner (80% strongly agreed/agreed—“I do both short-term and long-term planning”; Fayol, 1916; Gulick, 1937).

Strong participant responses to the strategic roles were surprising in light of the suggestions that CEOs do not have time for strategic planning (Carlson, 1951; Mintzberg, 1973) but may suggest a shift in the CEO role based on the challenges of difficult economic times. Many of the study participants referred to predepression and postdepression sales revenue numbers and employee numbers, suggesting that many of the organizations were deeply affected by the economic challenges of the past few years. Although the addition of chief strategy officers to top management teams has been recommended (Breene et al., 2007), economic realities may be forcing CEOs to fill this role themselves to insure survival of the organizations they lead. Timing of the survey may have also influenced responses to the strategic roles. The survey was sent in late October 2010, a time when many companies are finalizing projections for the current year and creating financial and operating strategies for the coming year. A strong focus on strategy at the time of taking the survey may have influenced responses to these roles.

The diplomacy role categories were more controversial in terms of agreement by the participants. Figurehead received the majority of strongly agreed/agreed responses in this category with 87.5% (“I represent the organization in formal matters”; Mintzberg, 1973). Liaison received 74.2% strongly agreed/agreed responses (“I interact with peers and others outside the organization to gain favors and information”; Mintzberg, 1973), and link received 66.9% strongly agreed/agreed responses (“I link the external world to the world inside the organization”; Lafley, 2009). The link role was described by Lafley, former CEO of Proctor and Gamble, in 2009 as the one role that only the CEO can perform. This new role may resonate more strongly with CEOs from large, global organizations like Proctor and Gamble but may have been less recognized by study participants because the majority of them (86.1%) were from privately held companies.

The operational roles were fairly controversial, yet organizer (“I make sure deadlines are met”; Fayol, 1916) and analyzer (“I focus on efficient management of the internal operating system in the interest of serving existing products/markets”; Hart & Quinn, 1993) received 65.6% and 60.8% strongly agreed/agreed responses, respectively. Controller (“I make sure projects are completed on time”; Fayol, 1916) received 55.1% strongly agreed/agreed responses while almost 30% of participants neither agreed nor disagreed that this is a CEO role. Operator (“I make sure that day-to-day operations are being completed in a satisfactory manner”; Howe, 1988) was even more controversial. Of the responses, 47.2% strongly agreed/agreed, 25.1% neither agreed nor disagreed, and 27.7% disagreed or strongly disagreed that operator is a CEO role. The controversy in these roles may be the result of organizations that employ chief operating officers (COOs) or other C-level executives responsible for operational activities. It is also possible that many CEOs do not feel it is their role to direct the day-to-day activities of their organizations and in many cases, they may not have time to do this. It may be more accurate to view the CEO role as one that guides the general direction of the organization from a 50,000 foot view, with much less involvement in day-to-day details.

The informational role category, which was changed substantially by factor analysis, was the most controversial. While negotiator (“I am compelled to enter negotiations on behalf of my organization”; Mintzberg, 1973) received 65.4% strongly agreed/agreed responses, the remaining roles received less than 50% strongly agreed/agreed responses. Of responses to the spokesperson role (“I disseminate the organization’s information into the business world”; Mintzberg, 1973), 46.8% strongly agreed/agreed, 23.3% neither agreed nor disagreed, and 29.9% disagreed or strongly disagreed that this was a CEO role. Technical expert (“I am the expert on product and market”; Lau et al., 1980) was split as follows: 38.4% strongly agreed/agreed, 32.4% neither agreed nor disagreed, and 29.2% disagreed/strongly disagreed with this role. Finally, problem solver (“I am the person who solves the organization’s problems”; Lau et al., 1980) was split as follows: 36.1% strongly agreed/agreed, 39.1% neither agreed/disagreed, and 24.8% of participants disagreed/strongly disagreed that this is a CEO role.

In today’s business environment, the negotiator role may be delegated to corporate attorneys more qualified to navigate the complexities of legal negotiations. This may be the result of a much more litigious and complicated business environment today than during the 1970s when Mintzberg identified this as a CEO role. The spokesperson role is frequently the responsibility of the investor or public relations department and this may partially explain disagreement with this role. The roles of technical expert and problem solver are likely delegated to individuals more familiar with the day-to-day operations of an organization, making these roles more controversial among participants. In high-technology and biotechnology firms, engineers and scientists are frequently the experts in the technical aspects of the business, whereas CEOs are viewed as the experts in business management and leadership.

Factor analysis results indicated that four roles did not load in this category and two that loaded together were not used in the final analysis of data. Four of these roles, monitor (“I receive and collect information enabling me to develop a thorough understanding of my organization”; Mintzberg, 1973), disseminator (“I transmit special information into the organization”; Mintzberg, 1973), entrepreneur (“I initiate changes within the organization”; Mintzberg, 1973), and consultant (“I provide advice on issues that arise within the organization”; Lafley, 2009) received strongly agreed/agreed responses in excess of 90%. Conflict handler (“I handle conflicts that arise between individuals and outside organizations”; Castaldi, 1986), and commander (“I give orders to employees”; Gulick, 1937), received strongly agreed/agreed responses from 76.6% and 62.0% of participants, respectively.

In summary, all 31 roles included on the survey received some strongly agreed/agreed responses and all 31 roles also received neither agreed nor disagreed, disagreed, and strongly disagreed responses. The leader role was the most agreed-on role and problem solver was the least agreed on by the study participants. The interpersonal role categories were most agreed-on by all participants, whereas the informational role categories were least agreed on. Many of the CEO roles identified in the 20th century still reflect the role of the CEO in the 21st century, even though the business environment has changed significantly. In addition, new roles have been identified, confirming the complex and evolving nature of the role of CEO (Lafley, 2009).

Research Question 2

Research Question 2 asked, “How much time do CEOs approximate they spend in six categories of roles? In direct conflict with Mintzberg’s theory, the participants indicated they spend 21.0% of their time in the strategic roles. Mintzberg (1973) argued that CEOs do not have time for strategy. There are many possible reasons for the substantial amount of time study participants allocated to the strategic roles. It is possible, but hopefully unlikely, that the responses were based on how CEOs think they should allocate their time. It is also possible that the responses had something to do with the timing of this survey. The survey was sent late October 2010 through late November 2010, which is the time when many companies are finalizing current year projections and creating budgets and strategic plans for the following fiscal year. The participants may have been spending more time focused on strategy when they received this particular survey, influencing their responses. It is also possible that in light of the current economic conditions, CEOs need to be and indeed are spending more focused time on strategy. It is possible that the complexity and uncertainty of the business environment today has forced CEOs to finally focus their time on strategy while leaving less urgent roles, like operations, to COOs, chief financial officers (CFOs), and chief information officers (CIOs).

Interpersonal, decisional, and operational roles received almost equal allocations of time, 17.5%, 18.1%, and 18.0%, respectively. One might expect to find more time spent in the interpersonal roles (leader, motivator, and director) based on the strong agreement with this role category. Conversely, the controversy within the operational roles would suggest less time might be spent in this category; however, this is not the case. Further exploration of this inconsistency might be best studied using interviews or other qualitative methods to understand why CEOs spend time in roles they may not necessarily believe to be their roles.

Participants spent 14.6% of their time in the informational roles. This may be due to the use of investor relations and public relations departments today that might not have been used 40 years ago. Participants spent the least amount of time in diplomacy roles (10.8%), which included link, liaison, and figurehead. The link role did not resonate strongly with participants, perhaps, because there were more privately held company CEOs than publicly traded company CEOs in this sample.

Research Question 3

Research Question 3 asked, “What roles are identified by CEOs that were not identified in the literature on the role of CEO?” Other roles suggested by participants that are different from the roles included on the survey include coach, mentor, cheerleader, succession planner, consensus builder, liaison with board of directors, community involvement and outreach, driver of sustainability, investor relations, financial oversight, fund raising, enabler, chief storyteller, recruiter, thought leader, and industry positioner. The chief storyteller role was also suggested in the pilot study of the survey and included the description “telling stories to reinforce the culture.” The board of directors liaison role was mentioned by several participants as a key role of the CEO, yet participants did not feel the survey reflected this important role.

It is not surprising that the CEO’s role as liaison with the board of directors was mentioned. In light of scandals like Enron, boards are now held to higher levels of accountability than ever before and may be more involved with the CEO than they were prior to Sarbannes Oxley (2002). The current focus on corporate social responsibility and increased regulatory pressure created by the Dodd-Frank Wall Street Reform and Consumer Protection Act (the Dodd-Frank Act), intended to enhance corporate governance, market structure, and corporate disclosure (BDO Knows: SEC, 2011; Wood, 2011), may increase the difficultly in accessing CEOs from publicly traded companies. The extensive rule making and regulation forced on public companies certainly may influence the complexity of the CEO role, in addition to putting the CEO in a defensive posture.

The new roles identified by participants are important for future development of the survey instrument and for theory building or theory refinement related to the role of CEO. Descriptions of these roles are necessary for refining the survey instrument and may be best obtained through a focus group of CEOs.

Hypothesis 1

Hypothesis 1 was partially supported by the findings. There were no statistically significant differences between male and female CEOs in terms of time spent in the six role categories. Males and females showed statistically significant differences in agreement with the operational, strategic, and diplomacy role categories; however, effect sizes were small to medium, d = .23, .29, and .39, respectively, leading to the conclusion that there is very little practical significance in the response differences between male and female CEOs. In all cases, the female participants agreed more strongly with the roles than their male counterparts, and female participants agreed more strongly with all of the roles than male participants.

Most empirical research on female CEOs focuses on the glass ceiling or on differences in compensation structure between males and females. The focus of this study, CEO roles and time allocations to roles does not appear in current research on CEOs. The findings of this research suggest that male and female CEOs may not be substantially different in terms of how they execute the role of CEO. It seems logical that female CEOs would need to know the role as well as their male counterparts. This is further supported by Hypotheses 1(a) and 1(b).

Hypotheses 1(a) and 1(b)

Female CEOs will spend more time in interpersonal roles than their male counterparts and female CEOs will spend less time in operational roles than their male counterparts were supported by the evidence, but the differences were minimal. It seems there are only minor differences in role agreement and time allocated to roles between male and female CEOs and this is not surprising. Female CEOs may hold the CEO position because they understand the expectations of the role just as well as their male counterparts.

Hypothesis 2

Hypothesis 2 was not supported by the data. CEOs from public companies were not substantially different than CEOs from private companies in their time allocations. This conflicts with the belief that the ownership status of a company affects the role of the CEO (Mintzberg, 1973). It seems that the impact of publicly traded ownership may not be as significant today as it was in the past. More likely, today private company CEOs experience similar pressures as their public company counterparts, a result of the complex business environment and strong competition.

Hypothesis 3

Hypothesis 3 was not supported. There were small differences in agreement with the informational roles between finance majors and technical/engineering majors, with the latter agreeing more strongly that these are CEO roles. There were also small differences in agreement with the decisional role category between operations majors and technical/engineering majors, with the latter agreeing more strongly that these are CEO roles. Overall, there was stronger agreement on the interpersonal and decisional role categories, with means ranging from 1.48 to 1.66 (1 = strongly agree) than the other categories. Time allocations were also different. Operations and technical/engineering majors had higher mean ranks in the operational role category than finance majors. In terms of practical significance, all of the effect sizes were small.

Hypothesis 4

Hypothesis 4 was not supported. There were statistically significant differences with small effect sizes in the interpersonal, operational, strategic, and diplomacy categories. In all categories, the CEOs with other C-level executives had higher mean ranks than those without other C-level executives, with the exception of the operational roles. This may indicate that CEOs who use other C-level executives need to spend less time in operational roles than those who do not.

Hypothesis 5

Hypothesis 5 was not supported by the data. These variables explained a very insignificant amount of the variance in role agreement and time allocations across the participants. This leads to a very important question: What factors do influence role agreement and time spent in roles by CEOs? This question remains unanswered by this research study. One could speculate that the contextual nature of business may influence this role including the economy, the business cycle, and the uniqueness of specific industries. Some have indicated that the CEO’s personality influences the role and the way time is allocated (Hart & Quinn, 1993; Zaccaro, 2001). It is also possible that multiple regression was not the appropriate model for understanding these relationships.

Hypothesis 6

Hypothesis 6 was not supported. There were small differences in role agreement in the informational, interpersonal, and diplomacy role categories. Public company CEOs agreed more strongly with the informational and diplomacy roles and private company CEOs agreed more strongly with the interpersonal roles. Although the differences were small, they seem logical. A CEO in a publicly traded company is more likely to be known by name than a CEO in a privately held company; therefore, the public company CEO may identify more strongly with the leader role due to his or her notoriety. Private company CEOs are less likely to have investor relations or public relations departments handling information requests; this may make the informational roles resonate more with private company CEOs. There were small differences in time allocations between the CEOs in the interpersonal role category.

Hypothesis 7

Hypothesis 7 was supported by the data. At the beginning of this research project, the researcher believed the most productive way to access CEOs would be through personal contacts (Martinko & Gardner, 1985; Thomas, 1995). The problem with this approach was the limited sample size. Thus, a database of companies that included CEO e-mail addresses was purchased. This random sample was much more fruitful than anticipated and led to a stronger study, yet there were no significant differences between the two groups of respondents.

Hypothesis 8

Hypothesis 8 was not supported by the data. In fact, there were statistically significant differences on agreement in all role categories across three levels of company size based on employee numbers, small, medium, and large. Company sizes were defined by the World Bank standards of small (less than 100 employees), medium (between 100 and 500 employees), and large (more than 500 employees; www.worldbank.com).

There were differences in time allocations between small- and medium-size companies in the informational (medium effect size), operational, decisional, strategic, and diplomacy role categories (small effect size). Small- and large-size companies differed in the interpersonal (medium effect size), informational, operational, and strategic categories (small effect sizes). Medium- and large-size companies differed in the interpersonal and decisional categories (small effect sizes). These differences lend support to research on CEO failure rates and the thesis that “something changes when a company reaches a certain size that makes it somehow different to manage” (Tuck & Earle, 1996, p. 19).

Agreement with Roles

Mintzberg’s (1973) theory about the role of CEOs was a foundational theory for the conduct of this research. All ten of the roles identified by Mintzberg received support by the study participants as roles of the CEO. Factor analysis indicated that three of the roles, monitor, disseminator, and entrepreneur, may not be measuring what they were intended to measure. Two roles, spokesperson and negotiator, were controversial roles from the CEO’s perspectives, perhaps, because these roles can be delegated to investor relations and legal departments. The data revealed that the informational role category may not be as relevant today as it was during the 1970s.

Although Mintzberg (1973) disagreed with Gulick (1937) and Fayol (1916) about the practice of management as science, their seven roles, commander, director, staffer, organizer, controller, coordinator, and planner, still received substantial support from CEOs as roles that are reflective of the CEO. Four roles suggested by Hart and Quinn in 1993 received support as being roles of a CEO: motivator, taskmaster, analyzer, and vision setter. The strategic roles of innovator and transformer (Galambos, 1995), strategist (Stata, 1988), and creator/maintainer of culture (Sashkin & Fulmer, 1988) also received support as CEO roles. Link and consultant (Lafley, 2009) were supported as CEO roles, along with conflict handler (Zaccaro, 2001). Roles receiving less support included technical expert (Lau et al., 1980), problem solver (Zaccaro, 2001), and operator (Howe, 1988). Of the 31 roles suggested by the literature review and used in the survey instrument, 26 roles received strong support as being CEO roles from the study participants.

Though not the primary purpose of this research, it now seems possible to begin the process of updating and refining theory that informs the role of CEO. Roles that no longer resonate with CEOs, like technical expert, problem solver, and operator, may now be considered the roles of other C-level executives within an organization. New roles including board of director liaison and chief embodiment of the values of the organization can be added as roles reflective of the changed nature of the CEO role.

Time Allocated to the Role Categories

This study is believed to be one of the first to ask CEOs to allocate their time across role categories. Some researchers have attempted to understand how CEOs spend their time while observing them at work, whereas others have asked CEOs about the importance of the ten roles identified by Mintzberg (Beggs & Doolittle, 1988; Mintzberg, 1973; Tengblad, 2006). It was surprising to find 21% of participant’s time was spent in strategic roles and 18% of their time was spent in operational roles. The operational roles were not strongly agreed on as the role of CEO. Table 1 provides a summary of the time allocations from this research compared with other studies on CEOs that provided time-related data.

OPEN IN VIEWER

Table 1

Comparison Data: Time Allocations to Role Categories

	Mintzberg (1973)	Whitely (1978)	Tengblad (2006)	Glick (2011)
Role categories	N = 5	N = 7	N = 4	N = 1,202
Informational	40.0%	65.9% (includes interpersonal)	61.0%	14.6%
Interpersonal	39.0%		29.0%	17.5%
Decisional	21.0%	34.1%	10.0% (includes strategy)	18.1%
Operational	—	—	—	18.0%
Strategic	—	—	—	21.0%
Diplomacy	—	—	—	10.8%

Factors Influencing Role Agreement and Time Spent in Role Categories

Gender, major of last degree or industry focus, having other C-level executives working for them, and company ownership status, that is, publicly versus privately held, had only small influences on agreement with CEO roles and time allocations. When combined, neither age, years in current job, years as a CEO, last degree earned, and company size in sales revenue were predictive in determining agreement with the CEO roles nor was there much association between these variables and role agreement. Company size in terms of employee numbers was a factor affecting the most significant differences in the way CEOs responded to role agreement and how they allocated their time.

It is likely other factors may have more influence on the role of CEO than were considered by this study. These factors include current economic conditions, availability of capital, whether a company has just been sold or is in acquisition mode, and company stability (existing companies versus start-ups). Several CEOs in this study indicated that they were spending more time on financial issues, a direct result of the difficult economic conditions they face. Some CEOs indicated that managing the board of directors took a substantial amount of their time yet did not see this role reflected in the categories presented in the survey.

Limitations of the Study

There were many factors that limited this study on CEOs. The database purchased was done so based on the availability of CEO e-mail addresses; however, many companies are still able to keep this information confidential. In some cases, the e-mail addresses included info as the name, so these went to a general e-mail address that most likely never reached the CEO. Many CEOs have their e-mail screened by assistants and it is possible many CEOs never saw the survey. Although there were a wide variety of company sizes and industries, private companies outnumbered public companies six to one. As expected, male CEOs outnumbered female CEOs seven to one.

There are many control variables that may have provided additional depth to this study. Industry-specific variables not controlled for in this study include geographic region(s) of the organization, industry sector, products imported or exported by the organization, whether the organization is service oriented, organizational financial health, business life-cycle placement and age of organization, and organizational culture, that is, whether it is a hierarchical or team-based organization. Individual CEO characteristic variables that were not controlled for in this study include schools attended, roles previously held, and number of job or industry changes. Additional research that controls for these organizational or individual-specific variables may provide additional new insight into the role of CEO and the way CEOs choose to spend their time.

Many of the comments received by CEOs indicated that the roles on the survey were too limiting and did not cover the breadth of roles that CEOs are expected to engage in. One CEO indicated that if they were doing an optimal job, they would be doing all of the roles included on the survey. It is impossible to know whether the participants answered the questions honestly or provided the answers that they believed were most favorable. There is also no way to know how accurate the time allocations were. The study was conducted only from the CEO’s perspective. The ability of participants to “self-select” can lead to errors including undermining “the extent to which the composition of the panel mirrors the composition of the general public and likely leads to biased results” (Dillman, Smyth, & Christian, 2009, p. 338). The expectations of boards of directors and other C-level executives were not considered for this study. This study was limited to CEOs in the United States, so the results may not be indicative of CEOs from other countries. Use of the survey instrument was a pilot study—it was designed specifically for this research, limiting its validity and reliability (Dillman et al., 2009). Other than the data obtained through EFA, there are no other measures available to support the instrument’s validity and reliability (Gliner et al., 2009).

Implications for Theory

One goal for this study was further refinement of Mintzberg’s theory of the role of CEO. According to Lynham (2002), theory requires an ongoing process of refinement and development which “ensures that the relevance and rigor of the theory are continuously attended to and improved on by theorists through further inquiry and application in the real world” (p. 234). Mintzberg’s theory was built deductively using a “research-to-theory” strategy (Lynham, 2002, p. 225). During the 1970s and early 1980s, many individuals contributed to the theory by using “theory-to-research” strategies, yet many of these theories were not proven to be useful in practice, thus leaving a gap that was the focus of this study (Lynham, 2002, p. 225).

Using the general method of theory building described by Lynham (2002), the ongoing process of theory development requires five phases, including conceptual development (the theory was originally conceived by Mintzberg in 1973), operationalization, confirmation or disconfirmation, and application. The operationalization phase requires testing in real-world situations. Many doctoral students attempted this phase leading to conflicting results, perhaps the result of small sample sizes. Confirmation or disconfirmation is the intentional design of research for this purpose, that is, to confirm or disconfirm a theory. This particular study has provided data that both confirm and disconfirm Mintzberg’s original theory. It has also provided confirmation and disconfirmation by roles proposed by other researchers, hopefully expanding and updating Mintzberg’s original theory about the role of CEO. A next phase for this research is the application phase in which a real-world CEO would test the effectiveness of the time allocations and the role categories within their organization. Finally, all theories require ongoing refinement and development to ensure their applicability to real-world situations (Lynham, 2002).

Implications for Research

Data collected during the survey can be used to further this study in several ways. One of the unanswered questions resulting from this study is “what factors do influence role agreement and time spent in roles by CEOs?” While company size based on employee numbers resulted in statistically significant differences in responses to role agreement and time allocations, is there additional influence based on industry? Is there a relationship between industry, company size, and agreement with roles and time allocations to role categories? Although not part of my original research, these findings could be important for CEOs in specific industry categories. This information could be particularly important for a CEO switching jobs in a new industry where roles and time allocations may be significantly different than those in their current job. Perhaps, this information could help CEOs transition more efficiently, helping to reduce excessive turnover.

The unsolicited qualitative data collected were not analyzed for this study. The unsolicited insights, remarks, criticisms, and words of encouragement were some of the most interesting data obtained, yet they were not consistent with the quantitative/postpositivist methodological strategy of this study. To deeply understand the factors influencing role and time allocations, in-depth discussions with CEOs are necessary. Although the hard, quantitative data are difficult to argue with, it is impossible to delve deeper for more meaning and understanding. For example, although the data showed little differences between male and female CEOs, intuitively it seems there must be interesting and insightful differences in the experiences and expectations of both male and female CEOs. These experiences can only be learned through qualitative methods. Research on the role of CEO has been criticized for its failure to find similarities while only looking for differences (Hales, 1999). The focus on differences may be due to the nature of hypothesis testing. Perhaps, qualitative methods would be more appropriate for identifying similarities in the role of CEO.

Implications for Practice

The results of this study could have practical implications for a variety of stakeholders including HR directors, boards of directors, CEOs, employees, investors, and consumers. A description of how implications for practice could affect these stakeholders is provided in the following section.