Those of us who teach in business schools rarely ponder the question, "Do managers matter?" Were we to do so, we would have to deal with, unsettling questions about the basic worth of our work, as well as the scruples of taking our students' time and money to help them become "better" managers. Perhaps we all implicitly have considered the issue of whether managers have much effect on organizational outcomes and have arrived at a reassuringly affirmative answer. After all, we are surrounded daily with news about executive brilliance and ineptitude, about CEOs saving companies and ruining companies, about shareholders and boards replacing ineffective top executives with promising new talent. "Of course managers matter," we say to ourselves.
However, not all who have carefully considered the issue agree, In fact, there is a school of thought, supported by some evidence, that top managers in general do not have much effect on organizational outcomes. Before proceeding too far with a book on strategic leadership, we must confront this fundamental issue.
Such is the purpose of this chapter. We start with a discussion of what it is that top executives do, tracing a relatively well-developed literature on executive roles, responsibilities, and arenas of action. We then turn to the central debate, first reviewing the arguments and evidence of those who are skeptical about managerial effects, then the perspective of those who argue that top managers have considerable influence on their organizations. Our resolution of the debate is not to pick one view as correct, but rather to propose a middle ground: sometimes managers matter a great deal, sometimes not at all; usually their influence falls somewhere in between. "Managerial discretion," or latitude of action, is the theoretical fulcrum we propose as a way of reconciling the two opposing camps. We then discuss the tendency for observers—and for society in general—to overattribute organizational outcomes to top executives, creating heroes and villains in the process and generally complicating the task of
objectively tracing managerial effects. We close the chapter with an inventory of research priorities.
What Do Top Executives Do?
The head of an enterprise, say a CEO or a division president, has numerous roles to fulfill, not all of which square with typical images of top executives at work. Classic conceptions of the CEO depict a big person behind a big desk engaged in big actions—planning, organizing, coordinating, commanding, and controlling (Fayol 1949]. Even loftier imagery is provided by Barnard (1938) and Selznick (1957), who emphasized the top executive's job as defining institutional mission and goals, maintaining institutional integrity, and obtaining cooperation from organizational members. Adding further to the picture of the remoteness of the CEO job was the post-World War II proliferation of analytic approaches for rational decision making: for example, operations research, formal long-range planning, and portfolio analysis. The joint emergence of computer technology and the professional!zation of management led to a belief, or heightened an existing one, that CEOs were, first and foremost, careful and comprehensive deciders of major courses of action.
It is precisely because of these entrenched beliefs about and images of top executives that Henry Mintzberg's book The Nature of Managerial Work (1973) was so startling and important. Mintzberg studied the minute-by-minute activities of five:experienced CEOs, each for a week. What he found was that CEOs are not buffered from daily minutiae and crises, they do not engage in much reflective planning, and decision making is but a modest portion of what they do. Instead, CEOs were found to work at a hectic and unrelenting pace on a wide array of tasks; their activity is characterized by brevity, fragmentation, and interruption; they gravitate toward the current and weii-specified and away from the distant and vague; they are attracted to and place credence in oral media; and they spend a great deal of time interacting—talking, cajoling, soothing, selling, listening, and nodding—with a wide array of parties inside and outside the organization.
On the basts of his data, Mintzberg distilled a set of ten managerial roles that he placed in three broad categories: interpersonal (figurehead, leader, and liaison), informational (monitor, disseminator, and spokesperson), and decisional (entrepreneur, disturbance handler, resource allocator, and negotiator). Table 2.1 presents a summary of Mintzberg's executive roles.
Some studies have found that Mintzberg's roles can be difficult to distinguish when observing discrete managerial activities (McCall and Segrist 1980; Kurke and Aldrich 1983). Other studies (typically examining various types of managers, not just CEOs) confirm the behaviors that Mintzberg observed, but argue that the roles can be further distilled, possibly down to as few as six: leader, spokesperson,
Table 2.1.   Summary of Mintzberg's Executive Roles
Role
Interpersonal
Figurehead
Leader
Liason
Informational
Monitor
Disseminator
Spokesperson
Decisional
Definition
Symbolic head; obligated to perform a number of routine duties of a
legal or social nature. Responsible for the motivation and activation of subordinates and for
staffing, training, and associated duties. Maintains self-developed network of outside contacts and informers
who provide favors and information
Seeks and receives a wide variety of special information (much of it current) to develop a thorough understanding of the organization and the environment; emerges as the nerve center of internal and external information of the organization.
Transmits information received from outsiders or from subordinates to members of the organization. Some information is factual, while some involves the interpretation and integration of diverse value positions of organizational influences.
Transmits information to outsiders on the organization's plans, policies, and actions, results, and so on; serves as an expert on the organization's industry
Entrepreneur
Disturbance Handler Resource Allocator
Negotiator
Searches the organization and its environment for opportunities and
initiates "improvement projects" to bring about changes; supervises
the design of certain projects as well. Responsible for corrective action when the organization faces
important, or unexpected disturbances. Responsible for the allocation of organizational resources of
all kinds—in effect, the making or approval of all significant
organizational decisions. Responsible for representing the organization during major
negotiations with others
Adapted from Mintberg 1973; and Pavett and Lau 19S2, 9
resource allocator, entrepreneur, environmental monitor, and liaison (Tsui 1984). Kotter's (1982] in-depth study of fifteen general managers confirmed Mintzberg's general portrayal of managerial work but concluded that it could be distilled even further: short- and long-term agenda setting, internal aridextemalnetwprk,|bjj|d-ingi and getting .the network toimplementthe agenda. None of these later studies is at odds with Mintzberg. When combined with yet other inquiries and models, they all indicate some basic dimensions of the top executive's job, which we now discuss.
Basic Dimensions of the Job
External and Internal Activities: Top executives operate at the boundary between their organization and the external environment (Thompson 1967). They gather information from outside, and they convey information, impressions, and reassurances to the outside. They alert insiders about external news and developments. They take actions to align the organization with the current and expected external environment (technology, market trends, regulatory forces, and competitors' initiatives); at times they try to modify the environment (through lobbying, trade associations, consortia, and joint ventures).
Strategy Formulation, Implementation, and Context Creation: Top executives may orchestrate the formulation of company strategy, including the choices of which products and markets to emphasize, how to outdo competitors, how fast to grow, and so on (Ansoff 1965; Porter 1980). Top executives also have a role in strategy implementation—allocating resources, establishing policies and programs, and developing an organization that is aligned with the strategic thrusts of the firm (Chandler 1962; Galbraith and Kazanjian 1986; Quinn 1980). And top executives create a context—through staffing, reward and measurement systems, culture and style—that influences the strategic choices made by the managers and technical specialists throughout the organization who are most familiar with marketplaces, technologies, and competitors (Bower 1970; Burgelman 1983).
Substance and Symbols: When we think of executive action, we usually gravitate to the substantive: acquiring or divesting a business, increasing a research and development (R&D) budget, opening a new factory, forming a task force to launch a total quality program, and so on. But, executives also operate in the world of symbols (Dandridge, Mitroff, and Joyce 1980). A symbol is something that has meaning beyond its inherent substance. By virtue of being at the top of the organizational hierarchy, executives' actions often convey extra meaning (Pfeffer 1981a). Some top executive actions are expressly symbolic, such as hosting a farewell dinner for a much-loved employee, holding a recognition ceremony to honor some extraordinary achievement, or personally appearing in the company's advertisements, However, to some extent, all executive actions carry added meaning, or what might be called ''symbolic1 fallout," conveying surplus messages to observers who are trying to detect the executive's intentions, values, predispositions, and where he or she is headed. Executive decisions, for example to promote one person but not another, to close one plant but not another, or to have an important meeting in a given location but not another, all convey meaning beyond their inherent substance to parties inside and possibly even outside the organization. In fact, some have said that the top executive's most important task is to establish and convey an "organizational meaning" (Barnard 1938). Despite its importance to management in general and executive leadership in particular, there has been little systematic research into the use of symbols by executives
43
(Armenakis et al. 1995; Smircich andStubbart 1985; Dutton andAshford 1993). Further, while there are noted exceptions (e.g., Gioia and Thomas 1996; Gioia et al, 1994; Westphal and Zajac 1998, 2001), empirical work is lacking.
Thus, top executives are engaged, at least potentially, in a wide array of roles, responsibilities, and activities. We say "potentially" because the roles that are emphasized vary immensely among executives. For example, the CEO of a publicly held corporation may engage in many more external activities (with security analysts, external board members, business journalists, and so on) than the CEO of a privately held company. The CEO of a company engaged in a turnaround effort will focus on different matters from the CEO of a company with abundant slack resources. In fact, Mintzberg (1973) laid out a series of descriptive hypotheses about how several contingency factors cause variation in managerial work. These contingency factors include environmental, organizational, situational, and individual factors, such as the executive's personality. We know of no studies that have attempted to directly test Mintzberg's hypotheses. However, as we will see throughout this book, numerous studies have documented the tendency for executives' own attributes (their experiences, education, functional background, personality, and so on) to affect their behaviors and choices,
That top executives would act on the basis of their own predispositions is fully understandable. Senior managers are embedded in ambiguity, complexity, and information overload. They encounter far more stimuli than they can comprehend, and those stimuli are typically vague, ill-formed, and contradictory (March and Simon 1958). Thus, the top executive faces the classic case of what the renowned psychologist Walter Mischel (1968) calls a "weak situation," that is, one in which the characteristics of the situation are not clear-cut enough to dictate a course of action. In such circumstances, the decision maker's personal frame of reference, not the objective characteristics of the situation, becomes the basis for action. It is precisely because of the multiplicity of executive roles, activities, and courses of action, along with the ambiguity and overload of the information confronting executives, that it is critically important to study how executives affect the form and fate of their organizations, Ultimately, executives' experiences, interpretations, and preferences greatly influence what happens to their companies.
Do Managers Matter? A Doubtful View
As intuitively reasonable as it may seem, the idea that top executives hold great sway over organizational outcomes is not universally held. Some theorists have set forth cogent arguments about the strict limits within which executives operate. And empirical evidence has been presented that, at least on its face, suggests top executives have far less effect on organizations than do other factors.
%0
Population ecologists particularly have argued that organizations—and their top managers—are largely inertial, hemmed in by environmental and organizational .constraints.. Hannan and Freeman (1977), for example, noted several internal constraints on managerial action: fixed investment in specialized assets, restricted information flows, internal political constraints, and entrenched norms and cultures. Similarly, they identified some significant external constraints; legal and fiscal barriers to entry and exit from markets, restricted access to external information, and legitimacy constraints.
Institutional theorists have argued that legitimacy constraints on organizations are particularly confining (e.g., DiMaggio and Powell 1983), Under great pressure to appear "normal" and rational, organizations must adopt numerous conventions that pull them into conformity with external expectations. Moreover, in the face of uncertainty, managers may be compelled to conclude that the least risky course of action is to imitate the choices of their counterparts (particularly the more successful ones) in other organizations. So, a process of "mimeticisomorphism* leads to remarkable homogeneity, particularly within an industry (Spender 1989; Hambrick, Geletkanycz, and Fredrickson 1993; Haveman 1993a).
An additional reason that managers may account for little variance in organizational outcomes is that managers as a group are exceedingly homogeneous (March and March 1977); that is, there is not much variance in the independent variable. Certainly on the surface, CEOs are not a diverse lot. In America's Fortune 500 companies, almost all CEOs are white men, aged fifty to sixty-five, who have college degrees and significant experience in large companies. In some countries, the pathways to large-company presidencies are even more restricted, often requiring graduation from one of a small set of elite universities (e.g., Kadushin 1995; Whitehill 1991; Kim and Cannella 2007). If top executives are drawn from a very narrow pool and then subjected to a long period of common socialization, we cannot expect them to exhibit much variety in thought or action.
Thus, for reasons of substantive constraint, institutional pressures for conformity and imitation, and extreme homogeneity of the top executive population, some have argued that managers do not matter. Several well-known empirical studies seem to point to that conclusion.
The most commonly cited evidence of minimal executive effects is Lieberson and O'Connor's (1972) study of top executives in large corporations. Using an analysis of variance procedure on a sample of 167 companies over a twenty-year period, the authors statistically isolated the portion of company performance (as measured by sales, profits, and return on sales) that could be attributed to the top executive in place in a given year. After the authors controlled for the year, industry, and specific company, leadership explained only between 6,5 and 14.5 percent of variance in the three performance measures examined. Lieberson and O'Connor concluded: "In short, all three performance variables are affected by forces beyond a leader's immediate control" (1972, 121).
1\
The second work often cited as evidence oi negligible managerial effects is Salancik and Pfeffer's [1977b) study of city mayors. Examining data on thirty U.S. cities over a seventeen-year period, the authors employed analysis similar to that of Lieberson and O'Connor. However, instead of explaining variance in organizational performance, Salancik and Pfeffer sought to explain variance in city expenditures in eight different budget categories. As did Lieberson and O'Connor, they inserted control variables, for city and year, before assessing the amount of variance explained by the mayor. Th^pfound'tnat the individual mayors accounted for 5 to 15 percent of variance in the expenditure categories. And, like Lieberson and O'Connor, they concluded that there is a relatively confined role for leaders; "Leadership in organizations operates within constraints deriving from internal structural and procedural factors and from external demands on the organization" (Salancik and Pfeffer 1977b, 492).
A more recent study points to a similar conclusion. From a thirty-year (1969-1999) sample of approximately 1,500 large public U.S. firms, Bertrand and Schoar (2003) generated a subsample of those senior executives (CEO, CFO, COO, and division presidents) who had worked in at least two firms during this time period. Controlling for year, industry, and firm-fixed effects, these authors identified—for several firm-level outcome variables—the proportion of variance attributable to CEOs and top management teams. Their results suggest that as little as 5 percent of variance in return on assets, for instance/maybe attributable to firms' top managers.
So, on the one hand, reasonable logic and large-sample data provide a basis for believing that top executives do not matter very much. And, while Lieberson and O'Connor (1972) and Salancik and Pfeffer (1977b) are relatively old studies, they are stil! cited as providing evidence that the "true" effect of leaders is small (Weber et al. 2001). On the other hand, a great deal of everyday observation, as well as other systematic studies, points to a very different conclusion.
Do Managers Matter? A Positive View
Some companies do not change much over time. But many do change, and at the hands of their top executives. Consider these firms; Nokia, the telecommunications equipment company; IBM,:,the information products and services firm; and Pearson, the media and publishing company. Over the last twenty years or so, in relatively short order, these companies have dramatically altered their mix of businesses. Their founders would not recognize them today, nor would their CEOs from even 1990. These companies are fundamentally different because of choices made by top executives,
Executives make many kinds of choices. Sometimes, as with the companies noted above, the choices are bold and quantum; sometimes they are incremental;
sometimes uiey maimam me aLams 41™, aumcuuico _.. but rather a failure to generate and consider choices. But managers act. As we shall argue throughout this book, they act on the basis of their own highly idiosyncratic experiences, repertoires, aspirations, knowledge of alternatives, and values.
Problems with Lieberson and O'Connor's Study
Before presenting affirmative evidence about managerial effects, we wish to return to Lieberson and O'Connor's (1972) oft-cited finding that top executives account for little variance in organizational performance. Their study, as influential as it is, had several methodological and analytic problems, all of which biased the results against observing managerial effects.
The most widely noted criticism of Lieberson and O'Connor's study deals with their dioice^^ (Hambrick and Mason 1984;
Romaneili andTushman 1988). Two of their three performance measures—saies and earnings—are primarily indicators of the firm's size. In their data analysis, the authors sought first to explain variance by using three independent variables: year, industry, and company. Not surprisingly, these variables were exceedingly strong predictors of sales and earnings, with explained variance as high as 97 percent. For example, if we know that a company is in the steel industry, and specifically it is U.S. Steel in the year 1950, our ability to estimate the company's sales level will be relatively high. However, only after controlling for industry and year was the analysis rerun with leadership—represented by a dummy variable for each of the individual CEOs—included to determine how much additional variance could be explained. Since by this point almost all the variance had been explained, the apparent added effect of leadership was nil. When Werner and Mahoney (1981) replicated Lieberson and O'Connor's study, they allowed the leadership variable to enter the analysis at the same stage as the other variables and found that leadership, or "stewardship," accounted for 44 percent of the variance in profitability of major firms.
Other problems bias Lieberson and O'Connor's study as well. First, they designated a new leader whenever a new president or board chairperson was appointed, without any attempt to identify the CEO per se. But if a chairperson (who is also the CEO) names a new president, there in fact has not been a change in CEO; if the president serves as the CEO and there is a change in the chairperson, there has not been a change in CEO; or if a chairperson relinquishes the CEO duties to an incumbent president, there is a change in CEO even though the two parties have not changed. In American companies, these are all common occurrences (Vancil 1987). Hence, Lieberson and O'Connor's method for assigning specific CEOs to particular periods of time must have contained considerable error, making doubtful any attempts to associate specific CEOs with performance levels in corresponding periods.
12
Next, Lieberson and O'Connor excluded from their sample any industries heavily populated with diverse firms, as well as any firms that engaged in major mergers or acquisitions during the period of the study. However, altering a firm's portfolio of businesses—through diversification, acquisitions, and divestitures—is the primary way for an executive to have an immediate quantum effect on the form and fate of the firm. By excluding such cases, Lieberson and O'Connor tightly restricted their sample to more incremental strategies and, not surprisingly, an apparently lessened executive effect.
Our point is not to dismiss the Lieberson and O' Connor study. The authors had good reasons for the research design choices they made. However, iheir chokes consistently biased their findings away from observing managerial influence on corporate outcomes. Hence, their study provides far less than the definitive word on the matter.
Evidence of Executive Effects
Beyond abundant anecdotal evidence that top executives can substantially alter organizations (e.g., Tichy andDevanna 1986; Tichy and Sherman 1993), numerous large-sample studies point to executive effects as well. Some of those studies, such as that of Weiner and Mahoney (1981), have been directly aimed at demonstrating the limitations of Lieberson and O'Connor's (1972) study, Other works have gone beyond methodological refinements, introducing important theoretical perspectives. For example, Smith, Carson, and Alexander (1984) used a sample of Methodist ministers to demonstrate that leaders who had been very effective in prior assignments tended to deliver higher performance in their current assignments (as measured by church attendance and financial statistics) than leaders who had been previously less effective. In their view, the inclusion of a measure of managerial quality enhances the ability to predict managerial effectiveness. Similarly, Pfeffer and Davis-Blake (1986), Cannella and Rowe (1995), and Rowe and colleagues (2005) found that the prior records of professional .sports team coaches helped predict their performance in new coaching assignments.
Numerous other studies have examined and found significant associations between executive attributes or succession and organizational performance. A few examples will serve to illustrate this abundant stream of research. Virany and Tushman (1986), for example, found that the management teams of better-performing microcomputer firms had significant prior experience in the industry and tended to include the firm's founder. Gupta and Govindarajan (1984) found that different types of general manager expertise were associated with business performance, depending on the strategy being pursued by the business. Murray (1989), from a sample of twenty-six oil companies, found that top management teams composed of members of diverse tenures outperformed those with more
homogeneous tenures. Haleblian and Finkelstein (1993) studied a sample of computer and natural gas companies and found that the size of the.tbp management team was positively associated with company performance, while a measure of CEO dominance was negatively associated with, performance.
NfoJiaU:..re^ direct effects on )
organ)zational perforinan«->. Some investigators have focused on understanding how top management characteristics are associated with strategies and structures, For example, Hage and Dewar (1973) found that the values held by top management teams affected their organizations' subsequent degree of innovation. Miller and Droge (1986) found that chief executive personality influenced the structure of the organization. And Helmich and Brown (1972) found that whether a new chief executive comes from inside or outside the organization affects how much organizational change will occur early in his or her tenure.
These comprise just a minor sampling of the evidence that managers have influence on their organizations' profiles and performance. We do not wish to imply that such influence is total or easy to exercise, but it exists,
Moreover, we do not seek to extol the virtues of top managers. Executives are worth studying as much for their limitations as for their achievements, In fact, population ecologists may have overstated their initial case against managerial effects, precisely because they required any such effects to be positive. This quote illustrates the ecologists' early view of the adaptation, or strategic choice, perspective: "According to the adaptation perspective, subunits of the organization, usually managers or dominant coalitions, scan the relevant environment for opportunities and threats, formulate strategic responses, and adjust organizational structure appropriately" (Hannan and Freeman 1977, 929; emphasis added).
Such an interpretation omits the possibility that managers scan the irrelevant environment and formulate responses inappropriately. Population ecologists tend to equate deteriorating organizational performance with an absence of managerial effect, when what they may be observing is simply unwise or unlucky choices on the part of managers. Part of the problem may be terminology. Ecologists use interchangeably the terms adaptation and strategic choice to describe the model that rivals theirs. Because adaptation clearly connotes success in adjusting to the environment, it may be that in observing that organizations regularly fail, ecologists assume that organizations do not adapt, and beyond that, that strategic choices are not made—or are made but not implemented.
Later work by populations ecologists envisioned a more significant role for top executives in influencing organizational outcomes (e.g., Hannan and Freeman 1984; Haveman 1992). In fact, some empirical research by ecologists has explicitly examined the effects of executive departures on survival rates of organizations (Carroll 1984; Haveman 1993b; Haveman and Khaire 2004). Such could not
0.^
have been considered under the earliest formulations of the ecological perspective. Indeed, more recently, strategic choice has been posed as a link or bridge between a number of diverse perspectives, as well as a key driving force behind more recent evolutionary perspectives (Child 1997). Viewing both organizations and the environments in which they are embedded as social structures, with numerous linkages between them, provides the underlying logic for this approach.
Managerial Discretion
So, do managers matter a great deal, all the time? No, the amount of leeway available to senior executives, even CEOs, varies widely. In an effort to bridge opposing views about how much effect top executives have on organizational outcomes, Hambrick and Finkelstein (1987) introduced and elaborated on the concept of executiyevdiscretiqri^ Depending on how much discretion
exists, an organization's form and fate may lie totally outside the control of its top managers, completely within their control, or more typically, somewhere in between.
For discretion to exist, an executive must have, and be aware of, multiple possible courses of action. As such, discretion is not absolute. It stems from contextual forces, but it also is derived from within the executive. Stated another way, one executive might create or detect alternative courses of action in a given situation, while another in the same situation might not be aware of such alternatives. Thus, as we discuss below, an executive's discretion is in part a function of his or her own characteristics, especially cognitive limits.
Moreover, an executive's discretion is rarely explicitly defined. Executives typically do not know exactly what actions might be allowed by powerful parties. So, they operate on the basis of rough estimates of the extent of their discretion, sometimes floating trial balloons to test the boundaries; occasionally they even overstep those boundaries, only then to be sanctioned by governing or powerful stakeholders.
A CEO's degree of discretion does not occur by happenstance, It is derived from three sets of factors: environmental, organizatibnllj and individual managerial characteristics. So, as stated by Hambrick and Finkelstein (1987, 379), "a chief executive's latitude of action is fundamentally a function of (1) the degree to which the environment allows variety and change, (2) the degree to which the organization itself is amenable to an array of possible actions and. empowers the chief executive to formulate and execute those actions, and (3) the degree to which the chief executive personally is able to envision or create multiple courses of action.'
As Figure 2.1 indicates, Hambrick and Finkelstein posited some specific determinants of discretion within each of these three spheres. We now discuss those.
Task Environment Product Differentiability +
Market Growth + --. litcrna! Organization
Industry Structure (especially oligopoly) Inertia! Forces (size, age,
Demand Instability + culture, capital intensity) -
Quasi-Legal Constraints Resource Availability +
Powerful Outside Forces Powerful Inside Forces
CHIEF EXECUTIVE OlSeRETlON
Managerial Characteristics
Aspiration Level + %x    Commitment to the Status Quo ~" Tolerance for Ambiguity + _ Cognitive Complexity + Internal Locus of Control ~L Power Base J-
Figure 2.1.   The Forces Affecting Chief Executive Discretion
Environmental Sources
The characteristics of the firm's task environment greatly affect the level of executive discretion and, in turn, how much influence managers have on organizational outcomes. Let us return briefly to Lieberson and O'Connor's study, which, as discussed above, is primarily known for demonstrating minimal managerial effects. A less-noted finding from their study is that managerial effects on corporate performance differed substantially across industries. Firms in the publishing and soaps/toiletries industries had the greatest amount of variance in profit margins explained by executive leadership, while firms in the clay products and shipbuilding industries had the least (see Wasserman, Nohria, and Anand 2001 for more recent analysis of cross-industry differences in CEO effects). Hambrick and Finkelstein (1987) attempted to explain and extend these results by arguing that the former industries provided far more executive discretion than the latter, and that, in general, environments confer discretion to the extent that (1) there is a relative absence of clear means-ends linkages, that is, where a wide range of options can meet stakeholders' nominal tests of plausibility; and (2) there is an absence of direct constraints.
ll-
In turn, Hambrick and Finkelstein set forth the following industry determinants of executive discretion:
® Product differentiability 0 Market growth
* Demand instability
0 Low capital intensity
• Monopolistic and purely competitive industry structures (as opposed to oligopolies)
® Absence of legal and quasi-legal constraints (e.g., regulation) e Absence of powerful outside forces (e.g., large, concentrated customers, suppliers, funding sources)
Preliminary attempts to identify high- and low-discretion industries, for purposes of empirical inquiry, relied primarily on qualitative application of Hambrick and Finkelstein's ideas. For instance, Finkelstein and Hambrick (1990) examined aggregate indicators of product differentiability, market growth, and so forth of sixteen major industries to select the computer, chemical, and natural-gas distribution, industries as high-, medium-, and low-discretion environments, respectively. Similarly, Hambrick, Geletkanycz, andFredrickson (1993) used qualitative, gestalt judgments to assign foods/beverages, computing equipment, and scientific/ measuring equipment as high-discretion industries, and public utilities and telecommunications services as low discretion, citing the wide differences between the two sets of industries in terms of differentiability, capital intensity, degree of regulation, and growth rates.
Subsequently, more rigorous approaches have been undertaken. Haleblian and Finkelstein (1993) used archival indicators of advertising intensity, research and development (R&D) intensity, market growth, and degree of regulation to create an overall index of discretion in the computer and natural gas industries, The components of the index were highly internally consistent, and the index score differed widely between the two industries.
Hambrick and Abrahamson (1995) used a panel of academic experts to rate the overall amount of managerial discretion in seventeen industries. They found a very high degree of reliability among raters, and, moreover, a high degree of agreement between the academics' ratings and those of security analysts who specialized in each of the seventeen industries. Hambrick and Abrahamson then examined the associations between the panelists' ratings and actual objective characteristics of industry discretion (from Compustat data), as originally set forth by Hambrick and Finkelstein. Using regression analysis, they were able to estimate the implicit weights that the panelists attached to specific industry characteristics (e.g., market growth) in rating an industry's overall discretion. The authors then applied these weightings of industry characteristics to determine the overall amount of discretion of fifty-three additional industries. Table 2.2 lists, in rank order of
Table 2.2,   Ratings of Managerial Discretion in Seventy Industries
Industry Name
Standard Industrial Code
Computer and software wholesaling Computer communication equipment Electro-medical apparatus Computer storage devices Perfume, cosmetic, toilet preparation Catalog, mail-order houses Medical laboratories 'Computer programming In vitro, in vivo diagnostics Help supply services ♦Motion picture production Photographic equipment and supplies 'Computer equipment Telephone and telegraph apparatus Variety stores
'Engineering/scientific instruments 'Games and toys
Computer integrated system design *Pharmaceuticals 'Surgical/medical instruments Women's misses, junior's outerwear Eating places
Miscellaneous amusement and recreation services
Industrial measurement instruments
Motor vehicles and car bodies
*Radio/TV communication equipment
Real estate investment trusts
Orthopedic, prosthetic, surgical appliances
State commercial banks
Newspaper publishing
Personal credit institutions
Chemicals and allied products
*Book publishing
Search and navigation systems
National commercial banks
Family clothing stores
Drug and proprietary stores
Women's clothing stores
Department stores
Electric lighting, wiring equipment
Television broadcast stations
Men's youth, boy's furnishings
Grocery and related products - wholesale
Converted paper, paperboard (except boxes)
Hotels, motels, tourist courts
on
S04S 3576 3845 3572 2744 5961 8071
7372 2835 7363 7313 3861 3570 3661 5331 3826 3944
7373 2834
3841 2330 5812 7990 3823 37 U 3663 6798
3842 6022 2711 6141 2800 2731 3812 6021 5651 5912 5621 5311 3640 4833 2320 5140 2670 7011
Discretion Score
6.89
6.72
6.72
6.62
6.60
6.44
6,43
6.38
6.36
6.16
6.08
5.99
5.77
5.70
5.66
5.63
5.55
5.55
5.54
5.42
5.32
5.22
5.21
5.19
5.18
5.17
5.15
5.07
5.06
5.06
5.04
5.02
4.92
4.91
4,81
4.79
4.78
4.75
4.75
4.73
4.72
4.72
4.71
4.68
4.67
(continued)
Table 2.2. [Continued)
Industry Name	Standard	Discretion
	Industrial Code	Score
Hazardous waste management	4955	4.65
semiconductors	3674	4.61
Insurance agents, brokers, and service	6411	4.54
Paper mills	2621	4.46
Engineering services	8711	4.46
Water transportation	4400	4.34
"Instruments to measure electricity	3825	4.33
Grocery stores	5411	4.32
Savings institutions, federally chartered	6035	4.32
*Security brokers	6211	4.27
Natural gas distribution	4924	4.05
Commercial printing	2750	4.03
Motor vehicle parts and accessories	3714	4.92
Air conditioning, heating, refrigeration equipment	3585	3.80
Phone communication (except radiotelephone)	4813	3.72
Railroads, line-haul operating	4011	3.51
Drilling oil and gas wells	1381	3.41
"Certified air transportation	4512	3.23
Petroleum refining	2911	3,07
Water supply	4941	3.04
'Trucking (except local)	4213	2.72
'Gold and silver ores	1040	2.42
*Petroleum/natural gas production	Oil	2.33
Electric services	4911	2.25
"Blast furnaces/steel mills	3312	2.08
Natural gas transmission	4922	2.01
* These seventeen industries were included in the set rated by academic experts and security analysts; multivariate analyses of objective industry characteristics provided the basis for rating the other fiftv-three industries
Adapted from Hambrick and Abrahamson 1995
discretion, the seventeen industries used for establishing the weights, as well as the other fifty-three. As can be seen, such industries as computer programming, perfumes and cosmetics, and motion picture production received very high discretion scores. Such industries as natural gas transmission, electric services, and water supply were rated as very low discretion.
Up to this point, most research examining environmental sources of discretion has equated a firm's environment with its industry. Recently, though, some work has begun to explore how discretion, may vary systematically at a national level. Using a size- and industry-matched sample, Crossland and Hambrick (2007) found that the proportion of variance in firm, performance attributable to CEOs was significantly greater in U.S. firms than in comparable German and Japanese firms.
These differences in CEO effects are consistent with prevailing cross-national differences in formal and informal national institutions [North 1990], such as legal tradition, firm ownership structure, board governance, and cultural values. Relatedly, Crossland (2007) used existing data on several national institutions to generate a rnulticountry taxonomy of discretion. This taxonomy. suggests^h^ systems (e.g., Anglo-American countries), te^ discretion than others (e.g., northern European .and -East Asian countries).
It is also possible that macro-environmental factors have brought about a general expansion of managerial discretion in recent years (Hambrick et al, 2004). Beyond the obvious trend of deregulation in many countries, more options simply exist on the organizational landscape. Companies can select unique combinations of businesses in which to be active; they can be fully active in a business or partly active through joint ventures or other alliances; they can select among myriad geographic locales for producing their products and still others for selling them; they can use full-time permanent employees or contingent temporary workers. In short,, societal and economic trends, as well as organizational innovations, have expanded the choices for senior executives, perhaps well beyond what existed when Lieberson and O'Connor (1972) conducted their study that pointed to limited managerial effects.
Organizational Sources
In addition to environmental factors, the organization may have characteristics that enhance or, conversely, limit the chief executive's discretion. These factors include inertial forces, such as organizational size, age, a strong culture, and capital intensity, all of which limit executive latitude. Large, mature organizations with very entrenched cultures are not easily changed. Their top executives operate under considerable inertial constraints.
Also affecting executive discretion is the amount of resources available to the organization, as well as internal political conditions (as determinedby the distribution of ownership, board composition and loyalties, and internal power concentrations). For example, executives have far more discretion when ownership is widely dispersed than when one or a few owners own concentrated blocks (McEachern. 1975; Hambrick and Finkelstein 1995). A CEO who is also chairperson of the board has more discretion than a CEO who does not hold both posts (especially when the chairperson is the prior CEO and strongly committed to existing policies) (Lorsch and Maclver 1989; Finkelstein and D'Aveni 1994; Harrison, Torres, and Kukalis 1988). In short, characteristics of the organization greatly affect how much latitude executives have over strategy and policy.
Recently, Shen and Cho (2005) developed a theoretical framework for explaining involuntary executive turnover, and their framework relied heavily on the discretion construct. Their treatment is particularly germane to the concept of
Id
34
discretion because they addressed several issues that had been missing from the broader literature. First, they pointed out that the management and economics literatures both use the term managerial discretion, but the term has very different meanings in the two fields. In the economics literature, managerial discretion describes the extent to which managers are free to pursue their own interests, rather than shareholders' objective of maximizing wealth (see, for example, Jung, Kim, and Stulz 1996; Williamson 1963). In the management literature, the concept of discretion refers to the range of options open to executives, while the extent to which those options represent the interests of executives versus shareholders is not expressly discussed. To resolve this contrast in usage of the same term, Shen and Cho divided the construct of managerial discretion into two dimensions; latitude of actions and latitude of objectives. Latitude of actions corresponds closely to the Hambrick and Finkelstein concept of discretion. Latitude of objectives, on the other hand, refers to the extent to which a manager can set objectives for the firm that may differ from those of owners. The authors then developed the logic as to how these two dimensions, independently and in interaction., influence the causal antecedents and performance outcomes of involuntary executive turnover.
Hendry (2002) also discussed the divergence between the discretion concept in the management and economics literatures, though not as directly as Shen and Cho. Hendry described the problem of "honest incompetence" and its implications for the agency relationship between managers and shareholders, or managers and other powerful governance forces. Because managerial competence is not assured, it becomes a second dimension that those involved in governance activities must consider.
Hendry's paper raises some very interesting dilemmas, highlighting the role of governance in developing managerial competence, and in ameliorating selection mistakes. The extent to which a manager is coached and developed may have a lot to do with his or her ultimate capabilities as a,strategic leaden And; the level of discretion is a key factor in this development process.
More recently, Finkelstein and Peteraf (2007) offered a new perspective on managerial discretion that posited a fourth source of discretion-^managerial activities. Building on research on agency theory and transaction costs, they argued that a focus on the activity level of analysis opens up the question of how managers might evade or minimize constraints imposed upon their actions. This leads naturally to a consideration of the dynamics of discretion, an aspect of the theory not yet addressed in the literature, but one that can offer insight on central organizational issues such as how managers enact environments, the nature of managerial capabilities, and the interplay of constraint and choice.
Finkelstein and Peteraf (2007) made three points. First, they argued that some types of activities circumscribe discretion more than others, just as some types of environments, organizations, and personal characteristics limit managerial
31
discretion more than others. Second, by focusing on the key attributes of activities—Finkelstein and Peteraf (2007) specifically highlight uncertainty, complexity, and observability—it is possible to develop predictions as to which types of activities offer more, or less, discretion, And third, managers can create or select activities in which they have greater opportunities to have an impact on organizational outcomes. In sum, Finkelstein and Peteraf (2007) offer a new perspective on discretion that extends the original conceptualization in interesting ways, specifically by bringing in the activity level of analysis to the core theory of discretion.
Individual Sources
As noted above, discretion is derived in part from executives themselves. By virtue of their personal characteristics, chief executives differ in the degree to which they generate and are aware of multiple courses of action. Some executives see alternatives that others do not. Some executives, because of their own persuasive and political skills, can consider options that others cannot. Hambrick and Finkelstein (1987) posited the following as specific individual-level attributes affecting discretion: aspiration level, tolerance for ambiguity, cognitive complexity, locus of control, power base, and political acumen.
So far, researchers have not, empirically examined these individual-level bases of discretion. However, this is a critically important arena for investigation, since the creation of discretion may be the critical ingredient in executive capability:
Managerial quality could be defined in part as the ability to perceive, create and enact discretion. Managerial excellence is a function of sheer awareness of options. Although it is an open (and researchable) issue, we suspect that managerial performance is more a matter of generating options than ofr^ftleGting^aj^ among a given set of options, to most
knowledgeable executives one will typically tend to stand out as the best. Thus, the opportunity for managerial contribution lies in improving on the list. (Hambrick and Finkelstein 1987, 374)
Namely, executives can shape their own discretion. Effective managers find and create options that others do not have. They may do this through creativity and insight, political acumen, persistence, or sheer will. Managers, even in a given situation, are not uniformly hemmed in. Child (1997), for example, noted that executives have various interpersonal linkages to the external environment, and can often use these linkages to influence the environment's effect on the organization, thus setting their own level of discretion (within limits).
Along those same lines, Carpenter and Golden (1997) provided evidence on the role of perception and locus of control1 in the discretion context. Their study
33
examined the age-old strategic leadership question, "Why do different managers, when confronted with the same situation, respond differently?"They argued (and empirically tested) for two distinct explanations. First, following Hambrick and Finkelstein (1987), they pointed out that an executive's locus of control will affect the amount of discretion the manager has: internals will have more discretion than externals. Second, the authors noted that managers can (and do) use impression management tendencies to influence how much power others perceive them to have. That is, through impression management, managers may be able to increase their own level of discretion.
Carpenter and Golden found qualified support for these ideas in a simulation involving executive MBA students. For example, they found that a given manager's perception of his or her own discretion relied upon locus of control, but only in low-discretion situations. Further, they found that impression management techniques could increase others' perceptions of the manager's discretion, but again, only in low-discretion situations. Their study provided important evidence that personality is an important factor in individual-level discretion, and that a given manager's perceived level of discretion is an important determinant of actual discretion.
Effects of Discretion
Executive discretion can be expected to affect a variety of phenomena of interest to organizational scholars. For example, Hambrick and Finkelstein (1987) argued that in situations of low discretion, the following could be expected: older CEOs who are promoted from within (to fulfill largely figurehead roles), low executive compensation, little use of incentive executive compensation, low administrative intensity, low involuntary turnover of CEOs, stable strategy, and changes in organizational performance tied closely to changes in the task environment. Situations of high discretion would tend to show opposite effects.
Equally important, however, is that discretion serves to attenuate the relationship between executive characteristics (values, experiences, and so on) and organizational outcomes. Namely, if high discretion exists, executive orientations become reflected in organizational, outcomes; if low discretion ..exjsts,:they do not. On this matter, research support is clear and consistent. For example, Finkelstein and Hambrick (1990) found that executive tenure was positively related to strategic persistence and strategic conformity to industry norms (reflecting presumably risk-averse and imitative tendencies of long-tenured executives) in high-discretion industries, but not in low-discretion industries. The authors also found that when the organization characteristics allowed top managers significant latitude—as indicated by abundant slack or small company size—strategic choices were more likely to reflect the tenure of the top executives than when slack was limited or the company was large.
In a similar vein, Haleblian and Finkelstein (1993) found that the relationship between TMT size and firm performance was significant in a high-discretion environment (computer industry) but not a low-discretion, environment (natural gas industry). Also, Forbes (2005) argued that small ventures are high-discretion settings (akin to Mischel's 1968 "weak situation") and therefore yield stronger links between executive characteristics and firm outcomes.
Abrahamson and Hambrick (1997) provided important evidence that discretion influences attentional homogeneity within an industry—and by imp! ication, the extent to which individual differences among managers will have effects on the decisions made. They developed an attention-interpretation-action framework to demonstrate that as attentional homogeneity increases among industry participants, the interpretations of strategic situations and the resulting actions decided upon become more and more consistent across participants. This study provided evidence to bolster the earlier assertion that executive choice sets may vary substantially over time, depending upon industry context.
Additional research, while not specifically invoking the concept of discretion, provides further evidence in line with the above suppositions. For example, Miller, Kets de Vries, and Toulouse (1982) found that CEO locus of control was strongly associated with organizational strategy and structure in small firms but not in large firms. The authors wrote, "these [small firms] might be more easily dominated than large ones, which, all things being equal, are more difficult to control" (page 249). In the same vein, Reinganum (1985a) found evidence that the stock market distinguishes between high- and low-discretion situations. On the announcement of CEO succession, stock prices rise abnormally, but only for small companies and when the predecessor CEO is totally departing the firm—these being conditions in which a new CEO can have an enhanced effect.
A good deal of research supports Hambrick and Finkelstein's ideas that discretion affects executive compensation arrangements, with executives in. low-discretion situations receiving relatively low levels of pay and little incentive pay. Rajagopalan and Finkelstein (1992) studied the electric utility industry from 1978 to 1987, a period of steadily increasing deregulation and, hence, increasing discretion. They found that executive compensation (for the CEO and top team) and the use of performance-contingent compensation increased over time as environmental discretion increased.
Rajagopalan (1997), using the same sample of electrical utility firms as mentioned above, described the importance of a fit between the level of discretion and the pay package provided to top executives. Using the Miles and Snow (1978) strategy categories of Prospector and Defender as proxies for the level of discretion, she argued that incentive compensation is an important determinant of firm performance only for Prospectors, because they naturally have higher discretion and therefore greater capacity for individual managers to influence performance.
35
Finkeistein and Boyd [1998] fleshed out the broad association between discretion, compensation, and firm performance. They predicted that compensation would be greater in high-discretion situations, to compensate managers for the fact that discretion makes the executive's job more complex, demanding, and risky. They further predicted that the relationship between discretion and compensation would be stronger in high-performing firms. Their results generally supported these hypotheses. Additionally, their manuscript discussed in detail some dimensions of organizational discretion and their measurement.
Two other studies also contribute to the discretion-incentive compensation predictions made by Hambrick and Finkeistein, Magnan and St. Onge (1997] provided evidence to support their hypotheses that the compensation-performance relationship is moderated by executive discretion. Their study involved 300 large commercial banks, and they developed some very interesting (albeit industry-specific) measures of discretion, Further, their results held across both accounting and market-based measures of performance, In a later study, St. Onge and colleagues (2001) qualitatively examined incentive plan effectiveness, using in-depth interviews with eighteen senior managers. Their results supported the notion that the effectiveness of stock option plans depends on the extent to which those targeted by the plans have the capacity (discretion) to take actions that directly influence stock prices.
Other studies, while not always explicitly investigating managerial discretion, have yielded corroborative findings. For example, a study by Kerr and Kren (1992), while not labeling firms as high or low on discretion, found that such indicators of discretion as R&D and advertising intensity strengthened the association between CEO pay and performance. Balkin and Gomez-Mejia (1987) found that high-technology firms, which tend to be characterized by greater levels of discretion (Hambrick and Abrahamson 1995), use incentive pay plans more than other firms do. And Napier and Smith (1987) found that the proportion of incentive pay was significantly greater in more diversified (and hence, higher discretion) firms. Further, Jensen and Murphy (1990b) found that the use of incentive compensation for CEOs was much greater in small firms than in large firms, prompting the authors to conclude: "Higher pay-performance sensitivities for small firms could reflect that CEOs are more influential in smaller companies" (p. 260).
A body of work in financial economics also provides insights into managerial discretion, The "investment opportunity set" is the full range of choices available to a firm or individual (Smith and Watts 1982). Similar to the managerial discretion literature, studies indicate that total compensation and the proportion of incentive-based compensation are higher when the investment opportunity set is greater (e.g., following industry deregulation) (Hubbard and Palia 1995).
A further stream of research suggests that executives may be aware, even if implicitly, of how much discretion they possess and that this awareness shapes their cognitive processing (Grinyer, Al-Bazzaz, and Yasai-Ardekani 1986; Javidan 1984).
3£
for example, in a large-sample study, Hambrick, Geletkanycz and Fredrickson (1993) found that in high-discretion industries, a firm's current level of performance was positively related to the top executive's commitment to the status quo (the belief that the organization's strategy and leadership characteristics in the future should remain as they are). This included, of course, the tendency for executives in poor-performing firms to believe that their firms should change. However, in low-discretion industries, no such association was found, leading the authors to state: "for the executive in a low-discretion situation, there is not a strong connection between current performance and a belief in the correctness of current organizational strategy and leadership profiles. In this instance, performance, be it high or low, emanates largely from uncontrollables—-the environment, the organization's confining history, etc." (1993, 406).
It is an open and interesting question as to whether executives modify their beliefs about the potency of executive action after sustained exposure to a higher low-discretion situation, or whether managers with certain types of beliefs and personalities (say, in terms of locus of control) are drawn to high- and low-discretion settings.
In general, executive discretion is an important construct for helping to bridge the debate about the influence of executives on organizational outcomes. Moreover, discretion may be a conceptual lever for improving our understanding of such matters as executive compensation, executive dismissal, organizational inertia, and executive personality,
Executive Job Demands
The concept of executive job demands refers to the "degree to which an executive experiences his or her job as difficult or challenging" (Hambrick, Finkeistein, and Mooney 2005, 473). While the concept of job demands has a long history in organizational behavior research and industrial organizational psychology (Xie and Johns 1995; Janssen 2001; Karasek 1979), it has received little attention as it relates to executive work. Most research on executives seems to assume that all executives experience their jobs as equally difficult. However, the extent to which an executive finds his or her work challenging is likely to have a number of observable outcomes,
Executive job demands arise from three classes of antecedents. The first, task challenges, refers to the general difficulty of the situation. Environmental hostility or munificence, competitive rivalry or stability, and the rate of environmental change all influence task challenges. Additionally, firm-level factors, such as the amount of resources or legitimacy, can greatly influence the level of task demands. The second antecedent of job demands is performance challenges. Most of these arise from external interests that may hold sway over the firm. Such challenges
3^
as siaKeiiumer pressures, concentrated ownership, and a viabie and active market for corporate control are important to the performance challenge dimension. Additionally, the firm's performance profile is a large factor in this antecedent. The third and final antecedent of job demands is executive ^aspirations. This factor includes such determinants as need for achievement, locus of control, and the extent of incentive alignment that the executive experiences in his or her salary structure.
Hambrick, Finkelstein, and Mooney (2005) proposed several associations between job demands and performance. For example, increased job demands are likely to lead to less strategic rationality in decision making, more reliance on past experience and repertoires, and more reflection of executive background in decision making. Paralleling the effects observed in more micro-oriented studies, the authors proposed that the overall effect of job demands on performance will be curvilinear, with higher performance occurring when executive job demands are moderate. The authors also proposed some interactions between, job demands, performance, and executive hubris. For example, executives who have performed well in extremely demanding situations may be more likely to develop hubris. Finally, the authors proposed some impression management behaviors that may follow from relatively high and relatively low job demand situations.
TSse Managerial Mystique
No discussion about whether managers matter would be complete without addressing the strong human tendency to believe that leaders matter. People seek to have heroes and villains as a way of explaining organizational and institutional successes and failures. Through the ages, people have blamed kings for droughts, prime ministers for poor economic conditions, and baseball managers for losing seasons. Humans gravitate to human (and simple) explanations for noteworthy events or trends. Indeed, this particular tendency is often referred to by psychologists as the "fundamental attribution error'' (Weber et al. 2001, 583),
The work of James Meindl and his associates has been particularly instrumental in enhancing our understanding of "the romance of leadership." In one paper, Meindl, Ehrlich, and Dukerich (1985) argue that leadership is a "perception" that allows people to make sense out of organizationally relevant phenomena. The authors explain the idea, that attributions to leaders will be greatest when organizational performance is extreme—-either very good or very bad. Their evidence, drawn from multiple methods and levels of analysis, is not definitive but is clearly intriguing. They find that business press headlines refer to a company's leadership in direct proportion to the company's performance: the better the performance, the more attention is showered on leaders. At a more macro level, the authors find
that the number of doctoral dissertations written on leadership subjects increases in bad economic times ("Where's the leadership to take care of this mess?") and that the number of articles in the business press dealing with leadership increases in good economic times ("Hurray for all this great leadership!"). (The difference between the pattern for dissertations and the press perhaps says something about the cynical lenses of academics.) Finally, in a series of laboratory studies, Meindl and his associates found that subjects, after reading a vignette, were relatively likely to ascribe extreme performance—either good or bad—to the leader of a business; more moderate or neutral performance was less likely to be attributed to the leader.
In a follow-up study, Chen and Meindl (1991) examined the role of the press in bestowing heroic and villainous status on leaders. Tracking the press accounts of the rise and fall of Donald Burr and People Express airline, the authors found that the press endowed Burr's ascendancy with a host of flattering images, then created an entirely new set of images of Burr to account for the company's collapse—all the while striving to demonstrate a consistency in the two distinct sets of portrayals. This project and the research stream it represents highlight the tendency of people—exacerbated, by the press—to attribute organizational outcomes to senior leaders.
Recently, in a series of experimental studies designed to rigorously rule out alternative explanations, subjects consistently misattributed poor performance to leadership, when the cause was very clearly due to group size (Weber et al 2001). Using some "weak link" games developed in game theory, the researchers explained how, for their experiment, group size was clearly and consistently the cause of poor outcomes. Whereas dyads nearly always come to optimal solutions in these games, groups of seven or more members seldom do so, regardless oi how well they understand how the. games work. In the experiment, the authors randomly assigned one group member as the "leader" and asked that person to exhort the group to behave according to very simple rules so that all would enjoy a favorable outcome. Very consistently, the larger groups failed to conform and tended to blame the failure on the group's leader. When given an opportunity, poorly performing groups tended to vote to replace the leader with another randomly chosen from among the group's membership. The authors argued that participants clearly understood how the game worked, but still failed to correctly attribute the outcome to group size (its true cause).
One interesting avenue for future research into executive attributions relates to several studies that indicate this "fundamental" attribution error (Tetlock 1985) may not be quite so fundamental after all (Harvey, Town, and Yarkin 1981). For example, Krull and colleagues (1999) found support for the idea that individuals in collectivistic cultures (such as China) were less likely to attribute outcomes to individuals, and more likely to perceive external causes, than were individuals from individualistic cultures (such as the United States). Thus, managerial
11
39
ijsj t%ft   '.ft'..«"'i'i lu mru i '-ii1 í
mystique, or CEO celebrity (Hayward, Rindova, and Pollock 2004), may in fact be culturally contingent.
Executives also generate their own attributions about their effects on their organizations. Here, the data are clear and quite consistent with the human tendency to manage impressions. Executives tend to take credit for Favorable outcomes and blame external forces for unfavorable outcomes. The predominant research method for detecting this pattern is content analysis of the letters to shareholders in annual reports (Bettman and Weitz 1983; Abrahamson and Park 1994). One such project captures the essence of the phenomenon in its title: "Strategy and the Weather.'1 Here, Bowman (1976) found that food companies that performed poorly very often blamed the weather and accompanying crop conditions, whereas food companies that performed well (and presumably faced the same weather) made no mention of the weather but instead pointed to the wisdom of their strategic choices.
The attributions made about executive influences on organizations are exceedingly interesting in their own right. Of course, these attributions also pose complications for the researcher who is interested in trying to objectively detect executive effects.
Conclusion
This chapter leads to a set of interconnected summary statements. First, senior executives operate in a wide array of spheres, encompassing substance and symbols, decisional and interpersonal roles, and external and internal activities. Further elaboration and development of the roles of human and social capital in the executive context are needed.
Second, there are numerous avenues by which top executives can influence organizational outcomes. Moreover, situated where they are (i.e., at the top), executives would seem to have the power to make things happen. Nevertheless, some research has concluded that top managers, including CEOs, do not have a strong effect on organizational performance. And although reservations can be raised about the analytic aspects of those studies, they cannot be entirely dismissed. Constraints on executives do exist, more so in some instances than in others. Executives sometimes have very little latitude of action, sometimes a great deal, and usually somewhere in between. Executive discretion is the concept that allows us to describe and understand how much leeway exists. Discretion stems from factors in the environment, the organization, and the executive him- or herself.
Third, further elaboration of the discretion construct, as well as its different dimensions in the agency and strategic contexts, is needed. While the extent to which powerful stakeholders afford executives latitude in setting firm-level
objectives is clearly important, it differs in important ways from the broader discretion construct that addresses the latitude of strategic actions available to the executive.
Fourth, even though executives rarely have total influence over what happens to their organizations, people tend to attribute extreme outcomes to leadership. This tendency gives rise to a romance of leadership—heroes, villains, and scapegoats. Executives themselves further complicate the.observer's ability to assign outcomes to die right sources by taking credit for favorable outcomes and pointing to "uncontrollable factors" for unfavorable outcomes.
Considerable research is still needed, not so much on the most basic elements of what managers do and whether they matter, but on how and when they matter. This need is particularly great since so much idealized imagery, prescriptive folklore, and naive attributions exist about top executives. Careful understanding of executives' roles and activities is warranted, Mintzberg (1973) and Rotter (1982) created a foundation for dissecting and classifying managerial roles, but too little research has extended these ideas. Particularly needed is an examination of the factors that affect an executive's involvement in various roles (external versus internal, decisional versus informational, and so on). Explanations based on environmental, organizational, temporal, and individual factors may allow important advances in understanding managerial work and even in generating prescriptions about "fitting" managers to specific circumstances. We particularly encourage research on the symbolic aspects of top executive work. We are convinced that this is an important side of executive behavior, yet very few systematic or general-izable insights about executive symbolism have been generated.
Executive discretion remains a fruitful target for research, Considerable work is needed in understanding the determinants of discretion. We particularly encourage examination of how organizational and individual characteristics affect the top executive's latitude of action, to complement the bit of progress made in understanding environmental sources of discretion (Hambrick and Abrahamson 1995).
Great opportunity also exists for research on the consequences of discretion. Some work has indicated that discretion affects executive compensation arrangements, but even here more needs to be known. Other possible consequences of discretion—including executive profiles, turnover rates, executive mobility and careers, administrative intensity, and executive personality—have gone largely unexplored (Rajagopalan and Datta 1996). Discretion,, we believe, will be an important theoretical fulcrum for understanding these and other important organizational phenomena.
Finally, one of the most promising areas of research will be executive images and attributions. How an executive is perceived obviously affects his or her own professional capital, but it also affects the firm's legitimacy and its ability to attract support from stakeholders. Executives no doubt engage in impression management to
/ A
improve their images; however, the press and other external information conduits {such as executive search firms and business associations] also greatly influence the ways in which executives are perceived. It may be that managerial attributions vary widely by national culture, with individualist countries such as the United States imbuing more of a managerial mystique than do countries with collectiv-ist cultures, such as Finland and Japan. We anticipate that executive reputation, stigma, prestige, and attributions will be prominent constructs in some of the most interesting research on top management over the next several years.
How Individual Differences Affec
Top executives operate in a world of ambiguity and complexity. Unlike convenient business school case studies, in which all the "reiev^it facts" arepacked into twenty-five pages, real strategic situations lack structure; the identification and diagnosis of problems are open to varying interpretations; and potentially pertinent information is often far-flung, elusive, cryptic, ^en contradictory. At odds with most strategy frameworks in textbooks, top executives do not deal in a world of tidily packaged, verifiable facts and trends. Effen if executives were able and inclined to conduct in-depth comprehensive analyses of their situations, they would typically arrive at widely differing conclusions, because strategic situations are not know-able, .riiey are only interpretable./
Consider, for instance, the Bfiyriad projections, estimates, and interpretations that entered into the decisiomoy Google's senior executives to acquire YouTube, an online video-sharing site, for a staggering Si.65 billion in October 2006. YouTube first launched its service in December 2005, a mere eleven months before it was acquired, and it had yet to make a profit. Moreover, because YouTube was privately held, any valuation of the firm would contain considerable guesswork and leaps of faith. Natut^lly, then, many critics panned Google's move. Concerns were raised about Goodie's increased exposure to copyright litigation, and some skeptics openly wondered how YouTube's grassroots business model could possibly ever yield a promt.
Presumably, other media companies such as Microsoft or Yahoo could have entered tru/fray, but decided that YouTube would never be a big moneymaker, or simply tMt$ 1.65 billion was too much to pay. Obviously, someone was wrong-— either tfie reluctant bystanders, or Google for paying so much. The actual payoffs in the/years ahead for. these parties will depend on dozens or even, hundreds of possible future events or trends—few of which can be estimated with any precision. No one knows what will happen, but that does not stop strategic decision
43