Power as an engine of influence and leadership has been of concern to priests, philosophers, and kings since the beginnings of civilization and to chiefs and medicine men before that. Who can influence whom clearly depends on who is more powerful and who is less so. Only with God on his side could Moses persuade the pharaoh to let his people go. Shakespeare’s plays are filled with concern about power and innocence, failing power, and personality and power (Hill, 1985). “To say a leader is preoccupied with power is like saying that a tennis player is preoccupied with making shots his opponent cannot return” (Gardner, 1986a, p. 5). This chapter explores the meaning of social power, the bases of power, and how power contributes to leadership in societies, communities, formal organizations, and small groups.1
Power is the force that can be applied to work. It is the rate at which energy can be absorbed. Social power is the ability to take actions and to initiate interactions. It is a force underlying social exchanges in which the dependent person in the exchange relationship has less power and the person with more power is able to obtain compliance with his or her wishes. Compliance implies acceptance of the more powerful person’s influence. It can be assent given enthusiastically or reluctantly. The compliant person depends on the more powerful person for desired outcomes that cannot be obtained from other sources (Emerson, 1964).
Power has been defined as the production of intended effects (Russell, 1938); the ability to apply force (Bierstedt, 1950); the right to prescribe behavioral patterns for others (Janda, 1960); and the intended, successful control of others (Wrong, 1968). But, as Gardner (1986a, p. 5) observed, “power does not need to be exercised to have its effect—as any hold-up man can tell you.” Most behavioral theorists maintain that power can be exerted without intention. Cartwright (1965) conceived of power in terms of controlling information and personal affection. Bierstedt (1950) focused on prestige.
Simon (1957) saw power as a manifestation of an asymmetry in the relationship between A and B. For J. R. P. French (1956), as well as Cartwright (1959a, 1959b), the power of A over B equals the maximum force that A can induce on B minus the maximum resisting force that B can mobilize in the opposing direction. For Dahl (1957), “A has power over B to the extent that A can get B to do something that B would not otherwise do.” For Pfeffer (1981a), power is A’s ability to change the course of action of B from what B would otherwise have done. A can overcome B’s resistance or reluctance to act. For Pres thus (1960), power was a matter of A’s and B’s rapport with each other. Emerson (1964) emphasized the dependency relationship between A and B. If B is more dependent on A than vice versa, B has less power than A. B’s dependence on A can be due to A’s control of resources that B cannot obtain elsewhere. It can be due to A’s ability to reduce uncertainties for B, or to A’s ability to cope with critical contingencies in the attainment of organizational goals based on the organization’s strategy, environment, or technology (Daft, 1983). A can obtain B’s compliance with A’s wishes. Thus Bagozzi and Phillips (1982) calculated that 64% of the power that suppliers have over wholesale distributors can be accounted for by the inability of the distributors to find other suppliers, the critical value of the resources controlled by the suppliers, the countervailing power of the distributors, and how much business is involved. According to Bennis and Nanus (1985, p. 17), power is “the basic energy needed to initiate and sustain action … the capacity to translate intention into reality and sustain it.” For Burns (1978), power can be wielded nakedly, as when people are treated as things; or it can be relational, collective, and purposeful.
Thibaut and Kelley (1959) regarded power as an exchange relationship in which one member has control over another’s behavior or fate. As Gardner (1986a, p. 5) said,
It is possible to think of the exercise of power as a kind of exchange. You want something from me and you have the power to produce in return certain outcomes that I want—or want to avoid. You can give me an A or flunk me. You can promote me to supervisor or reduce me to clerk. You can raise my salary or lower it. You can give or withhold love.
Individuals tend to maintain a balance in the exchange of social values (Homans, 1958). The differentiation of social roles that results in organizations is based initially on such exchange relations (Gouldner, 1960).
Blau (1964) and Adams and Romney (1959) saw power as negative in A’s use of deprivation, aversive stimulation, and sanctions to control B’s behavior. Harsanyi (1962a, 1962b) took a broader view and weighed social benefits against social costs in defining power relations. Bass (1960) defined the power of A and B as A’s control of what B needs and values. If B is satiated or uninterested in what A controls, A loses power over B. Burns (1978) agreed. Cartwright (1965) emphasized the importance of ecological control: power can be exercised by controlling resources or necessities, by occupying space, or by avoiding or boycotting a location. Thus a leader’s power is reflected in the ability to impose standards, limits, and boundaries on a group and to indicate what is expected of members to obtain rewards or avoid harm. It also hinges on the ability to enforce such rules. The boundaries are likely to be areas of increasing threat and impenetrability. This situation requires judgment by those who are being controlled about how far they can deviate without punishment (Timasheff, 1938). One way lack of power manifests itself is in hesitant speech: hedges like “I guess;” tag questions such as “Isn’t it?” at the end of an affirmative sentence; and disclaimers such as “Don’t get me wrong but …”. Powerlessness also shows itself in excuses and justifications for doing the wrong thing (Hackman & Johnson, 1991, p. 99)
The social exchange can be seen in the exchange relationship between the Roman patron and his freedmen clients and between the ward boss and his neighborhood constituents. The clients’ and constituents’ votes and obedience provided the bases of power for the patron and the ward boss, in return for which the followers received protection, security, and material support. Lobbyists can exert power over state and federal legislators through financial support of their political campaigns. The political leader in the United States disperses tangible divisible rewards, favors, contracts and services to his or her constituents. For example, Borowiec’s (1975) study of 83 Polish-American leaders in Buffalo indicated that they provided constituents with special information and personal assistance in dealing with public agencies. Privatization of government agency functions has meant billions of dollars in no-bid contracts for private firms in exchange for political support.
The exchange calculus may be complex; leaders of more powerful constituents are likely to be more powerful than leaders of less powerful followers. The leaders can mobilize the power of their followers and can be granted or deprived of power by their constituents. In combination with their constituents, they can collectively acquire more power (Burns, 1978). This mutuality in the power relation is seen when constituents must complete their assignments if they and the leader are to be rewarded. The slowdown, inattention, or corruption of constituents can result in a loss of benefits for the leader.
House (1984) operationalized the measurement of power by its effects. Power was equated with “the capacity to produce effects on others. These effects are achieved by the exercise of authority, expertise, political influence and charisma … each having a different base or source and each having different effects” (p. 26). But it is useful to maintain the distinction between power and influence, although the boundaries between power and influence remain unclear (Faucheux, 1984). Unfortunately, “power” is often used synonymously with “influence.” Tautologies have been endemic. We observe that A influences B; therefore, we conclude that A has more power than B. Inferences about power and its effects must begin with measures of power that are completely independent of observed relations between A and B. The observed relations are a product of the power differences between A and B, not the behavior observed. Leadership can be conceived of as the exercise of power (Berlew & Heller, 1983). Thus Milewicz (1983) concluded, after interviews with purchasing agents about their relations with their many suppliers, that the tendency of the agents to exert leadership in dealing with suppliers was a consequence of the agents’ self-perceived power. The agents had such power because they could, for example, control the length of the order cycle. Whether they fully used this power to exert leadership depended on their personal desire to be influential.
Leadership and influence obviously are a function of power. Power is the potential to influence. It is the probable rate and amount of influence of a person or the occupant of a position. Thus, for instance, Barber (1966) inferred that the power relations among members of a legislative committee could be accounted for by such factors as the rate of success in getting suggestions accepted and the rate of agreement received. Although Lord (1977) found that for 144 undergraduates, perceived leadership and social power were related, the unexplained variance required maintaining separate conceptual distinctions between them. Follert (1983) constructed a 20-item scale to measure the power of a manager. The scores obtained were related to the structuring leadership of the manager and to subordinates’ compliance with requests, but were clearly distinguishable from the scale of power.
We need to separate the holding of power because of: (1) one’s person, (2) one’s office, (3) the willingness to exercise it, and (4) the tendency actually to do so. Studies tend to confuse all four causes and sometimes even fail to distinguish between power and influence.
Status, gender, personality, and subordinates’ power moderate the effects of a leader’s power.
Status. How leaders use power will depend on their status. A leader higher in the organizational hierarchy uses more aspects of power to influence (Koslowsky & Schwarzwald, 1993). Vecchio and Sussman (1989) argued that the leader’s status altered the effect of his or her power on subordinate compliance. According to Fitness (2000), the status of a leader markedly altered reactions to a situation that provoked anger. Structured interviews were conducted with 175 participants about a work-related incident with a superior, peer, or subordinate that made the respondent angry. Angry superiors were more likely to vent their anger and consider the incident closed; angry subordinates were more likely to suppress their anger and consider the incident unresolved.
Gender. When men’s satisfaction with a heterosexual relationship decreases, they tend to become more coercive and unfeeling toward their partner; when women’s satisfaction with the relationship declines, they tend to fall back on their referent power based on being liked or identified with in the relationship (Schwartz wald, 1993).
Personality. Thomas, Dickson, and Bliese (2001) found that ROTC cadets’ extroversion partially mediated the relationship between the power of the cadets and their effectiveness as leaders. For 818 cadets, the direct-path correlation between their power and their leadership was only an insignificant .07. The indirect-path correlation from power to extroversion was .35; from extroversion to effective leadership it was .11.
Power of Subordinates. Power is not an absolute amount. As was already noted, it depends on the power of those to be influenced, both positively and negatively. The power of those who are influenced adds to the total power available in the situation and can be increased by the synergistic action of the leader and the followers. The leader’s power may be diminished to the extent that it can be offset by the power of individual followers or of the group led. Self-managing teams clearly reduce the power of a leader (Manz & Simms, 1980).
Bass and Valenzi (1973) developed one set of five-point scales to assess a boss’s power and another set of five-point scales to assess a subordinate’s power. The subordinate and boss each described how frequently the boss could override or veto any decisions made by the subordinate, grant or deny promotion or salary increases to the subordinate, reverse the subordinate’s priorities, control the size of the subordinate’s budget, and get support from a higher authority for what he or she wanted to do. In turn, the superior and subordinate each described how frequently the subordinate could bring outside pressure to support what he or she wanted; do the opposite of what the boss wanted him or her to do; maintain final control over his or her plans, assignments, and targets, regardless of what the boss thought about them; ignore the boss and submit requests to a higher authority; and nominate or vote for who would be the boss. The differences in the power of superiors and subordinates then became the metric of importance in predicting the superiors’ styles of leading the subordinates through, direction, consultation, negotiation, or the delegation of decision making (Chitayat & Venezia, 1984; Shapira, 1976).
Gardner (1986a, p. 16) pointed to the reversals that can occur between hierarchical rank in an organization and the power accompanying it: “Every experienced observer knows of cases in which such second and third level leadership is formidable—capable, for example, of paralyzing a newly appointed top executive.” Gardner concluded that the efforts to reform high school curricula in the 1960s failed because of the power of rank-and-file teachers whose support was not enlisted in the attempted improvements. The power of subordinates over their superiors is widespread when the skills and knowledge of the subordinates are hard to replace. On the one hand, superiors need to guard against becoming overly dependent on subordinates with power; on the other hand, the superiors need to maintain good interpersonal relations, trust, and openness of communications with them (Kotter, 1985a).
Numerous other factors reduce subordinates’ willingness to comply with the attempted influence of the power figure. For example, 240 Israeli subordinates of 40 police captains said they were less likely to comply with their captains if the captains coupled their power with acts of transactional instead of transformational leadership (Schwarzwald & Koslowsky, 1998).
As noted before, power can derive from one’s person or one’s position. Although it may seem otherwise, the evidence to date suggests that prospective followers tend to consider the personal power of a highly esteemed expert more important than the legitimacy and power to reward and punish that may derive from appointment to a position of leadership. Those with personal power can grant affection, consideration, sympathy, and recognition, and can secure relationships and attachments to others (Bass, 1960). For example, Rosen, Levinger, and Lippitt (1961) asked teachers and emotionally disturbed boys to rank in order of importance six items related to power. The two groups basically agreed on the relative importance of physical strength, sociability, expertness, fairness, fearlessness, and helpfulness as sources of power. But those with personal power were seen also as able to punish others by becoming more distant, formal, cold, and businesslike.
Bass, Wurster, and Alcock (1961) demonstrated that we want to be valued and esteemed mainly by those we value and esteem. We endow these people with personal power. Hurwitz, Zander, and Hymovitch (1953) showed that we seek the affection and support of those we hold to be personally powerful. Professional personnel were continually reshuffled to form 32 sextets. The power to influence and the extent to which each member was liked after the first half hour of interaction were assessed. Those with low esteem wanted most to be liked by those with high esteem. According to Yukl and Falbe (1991), personal power is based on expertise, referent power, esteem, persuasiveness, and charisma. Transformational leaders have more personal power and accordingly generate more compliance from subordinates than do leaders lacking in personal power (Atwater & Yammarino, 1996).
Identification and Personal Power. As will be described in Chapter 21, the charismatic leader can serve as the focus of positive emotional feelings, as the ideal object for psychoanalytic transference and identification (Kelman, 1958). The transference of coping with problems to parental figures provides followers with an escape from responsibility and making decisions—let the leader do it (Fromm, 1941). Followers identify with their successful superiors. Business executives see successful superiors as symbols of achievement (Henry 1949). Leaders of gangs of delinquents become the superego of the members. By submitting to the leader, members are relieved of all personal responsibility for their antisocial activities (Deutschberger, 1947).
The charismatic leader is the ultimate in personal power. Such a leader is personally endowed by followers with infallibility, wisdom, omniscience, virtue, and supernatural powers. Personal power is manifested in the emotional bonding between the leader and followers and in the followers’ dependence on the parental figure. They may have deep-seated affection for the leader. Charismatics such as the Reverend Jim Jones and Osama bin Laden have had the power to induce extremes of behavior—suicide and martyrdom—in their followers. Understanding such personal power requires an examination of strong emotional fixations that go beyond the ordinary considerations of social exchange, cognitions, rewards, and punishments.
The status associated with one’s position gives one power to influence those who are lower in status. Custom, tradition, rules, and regulations assign power to incumbents of positions. Some superiors can be the “purveyors of rewards and punishments” to their subordinates (Krech & Crutchfield, 1948). Leaders with positional power can recommend punishments and rewards, instruct group members on what to do, and correct each member’s job performance (Fiedler & Chemers, 1974). Such power can be wielded crudely or with subtlety. It may be obtrusive or unobtrusive and overt or covert (Ford, 1980). Overtly, a superior may have the power to recommend or deny pay increases to subordinates. Less obviously, a first-line supervisor, for example, may have power because of his or her influence with more powerful leaders who are higher up in the organization (Pelz, 1951). This power, in turn, may be due to the higher-ups’ attitudes about the role of supervisors who are lower in the organization, as well as to the personal attributes and merit of specific supervisors. Pelz found that only those supervisors who had influence with higher-ups exhibited behavior that made a difference in their subordinates’ satisfaction. Supervisors without influence “upstairs”—although they tried—were much less likely to have effects on their subordinates. Supervisors who had influence with higher-ups were much more likely to be able to obtain rewards for their employees. Their power was greater and, therefore, what they did in their own group made more of a difference to their subordinates.
Supervisors’ position in a hierarchy provides them with various sources of power. Supervisors command resources that are not ordinarily available to subordinates. They have discretion in assigning dirty and boring or interesting and challenging tasks. They can open or close doors to opportunities for the subordinates’ growth and advancement. They have power to reduce uncertainty or to prolong anxiety in subordinates by providing or withholding hard information. They can act as a buffer to keep their subordinates from becoming too vulnerable to an unstable external environment and thus sustain their subordinates’ optimism and hope (Hollander, 1978). According to Fiedler and Chemers (1974), leaders’ positional power combines with their esteem and orientation—task or relationship—to determine whether they will be effective in designated situations. According to Yukl and Falbe (1991), the power derived from one’s position is based on legitimacy, the power to reward or coerce, and access to and control of information. The use of positional power is more frequent among transactional than transformational leaders (Atwater & Yammarino, 1996).
Executives’ power over decisions depends on their functionally specialized roles. Accounting executives have more power over budgeting decisions; marketing executives have more power over decisions about advertising and distributing products. Extra power also accrues to those executives involved in activities critical to the firms’ strategic success, that is, to those who scan critical information and the environment for threats and opportunities (Hambrick, 1981). Similarly, although clinical interdisciplinary teams of health professionals are supposed to meet as equals, physicians dominate decisions about treatment, according to a survey of 137 team members’ descriptions of 19 meetings (Fiorelli, 1988).
One’s position in an organization can provide power because occupancy of the position gives control of important organizational resources and information. Control over resources such as the size of the staff, budgetary expenditures, and evaluation procedures may be involved (Gomez-Mejla, Page, & Tornow, 1982). A manager who controls such resources may be able to maintain a good working relationship with subordinates even if they have little trust in him or her (Novak & Graen, 1985). Power also accrues to positions that control critical contingencies originating in the environment or technology of the organization. This hinges on the extent to which the position makes it possible to cope with critical demands that the organization is facing (House, 1988b). The importance of positional power was illustrated when Geissler (1984) compared 131 women heads of university nursing programs with 108 women heads of other university departments. The heads of nursing felt significantly more powerful in their preferred roles than did the women who led other university departments. Position upgrades in a department were more likely to be recommended in job evaluations for those with more positional power (Wel-bourne & Trevor, 2000).
Socialization Processes. Members of organizations need to know how to use increases in their positional power (Taylor, 1986). Gouldner (1960) maintained that power is based on socializing norms, role differentiation, and organization that binds people together in the same social system. Gilman (1962) agreed that a power relationship is authoritative anywhere within the boundaries of a social system that gives it consensual support.
In many cultures, business transactions are conducted without written contracts because norms of obligation are regarded as binding (Macaulay, 1963). Similarly, politicians wield considerable power based on IOUs through which they can expect that favors they granted to other politicians in the past will be reciprocated when they call in their unwritten “chits” for services rendered.
Power Begets Power. House (1984) pointed out that individuals who are in positions of power not only can assert it successfully, but also can maintain and increase their level of power. Thus power-oriented individuals who gain positions of power will strive to retain and increase their power, since they are in a favored position to ensure that their power continues. Unless there are limits, incumbents of political offices tend to be reelected. (Over 99% of the members of the U.S. House of Representatives are reelected every two years.) On the basis of interview data from a study of the administrations of California governors Ronald Reagan and Jerry Brown, Biggart and Hamilton (1987) added the corollary that for people to sustain their power in an organizational setting, they must self-consciously exercise their power to signal those working for them to be aware of obligations to carry out assignments as expected.
Leaders may exercise both competitive and collective power. In a survey of 350 business and university managers, Roberts (1986) observed the widespread use of power to compete with bosses, peers, and subordinates, as well as the use of collective power to collaborate with others. McClelland (1975) saw the competitive exercise of power in the exertion of personal dominance or the search for victory over adversaries. This use of power was in contrast in cooperative settings with joining with others to form shared goals to instill in the collective a sense of power to pursue such goals. Undergraduates acting as managers in a simulated organization used their control of useful information to influence those acting as subordinates only if they saw themselves as being in a cooperative situation with their “subordinates” rather than as being in a competitive or individualistic situation (Tjosvold, 1985a).
The power of supervisors could be used in four ways, according to Jones (1983). (1) Power could be obtrusive or unobtrusive; (2) it could be situation-based or personal; (3) it could be professional or paternalistic; and (4) it could be exercised over process or output. Ford (1980) agreed on the utility of looking at covert power and showed experimentally that dependent participants in peripheral positions in a network will suppress decisions, particularly if they are uncertain about the preferences of the person in the central location in the network2 who has covert positional power.
Power over others depends on one’s control of resources to give or deny and control of rewards and punishments to distribute. Power manifests itself in influence over others (Bass, 1960). The labeling of positions may induce a sense of control of leaders over subordinates. Power may be sought for its own sake or for social reasons. The personalized leader is motivated to use power to dominate others; the socialized leader is motivated to use power to achieve mutually desired goals with others (McClelland, 1975).
Bass, Gaier, and Flint (1956) studied experimental male ROTC groups in which members varied in the amount of control they could exercise over each other’s avoidance of punishment—extra marching—because of demerits. Each member drew a card that specified whether he could exercise one, two, three, or four units of control to reduce the required marching. Members with four units of control attempted twice as much leadership as those with one unit of control under high motivation, but there was no difference in unmotivated groups with no demerits requiring the extra marching. Attempted leadership increased with the amount of control a member obtained in the lottery. But the amount of control ceased to have import when what was being controlled was unimportant to the member. Thus members attempted more leadership acts when they had a lot of control and when what was controlled was desired. Members had power to influence when they had control over what was desired (Bass, Pryer, Gaier, & Flint, 1958). Members’ attempts to lead increased with increases in their control and power, even if the control and power accrued from a lottery.
Influence is correlated with power but is not the same. Power is the ability to exert some amount of control over others. It correlates with authority and may be coercive in its effects. Influence is more likely to depend on persuasion, with followers having more latitude to accept or reject the leader. Since “the unfettered use of power can be highly dysfunctional in creating numerous points of resistance and lingering negative feelings” (Hollander, 1985, p. 489), leaders tend to prefer persuasion rather than coercion. Nonetheless, some members of a group tend to be more influential than others if they perceive themselves as having more power. Moreover, they tend to be more satisfied than members who have little power. They are also better liked, and their attempts to influence are better accepted. Lippitt, Polansky, Redl, and Rosen (1952) found that children at a summer camp who had more power attributed to them by other children were better liked, made more attempts to influence others, and scored higher in the initiation of behavioral contagion (imitation by others without intent or awareness). The camper’s perceptions of their own power were highly correlated with the power that others attributed to them. Consistent with these results, Levinger (1959) demonstrated that members of experimental groups, informed that they had more power than other members, tended to perceive themselves as actually having more power, became more assertive, and made more attempts to influence others.
Yukl (1998) noted that power has a variety of effects on influence. Power makes it possible for a leader to choose which influence tactics to use. For instance, the leader needs power to be able to make an exchange of benefits with a follower. A leader’s power may enhance or diminish the effectiveness of an influence tactic. For example, a leader with considerable power will influence decisions about relevant operations more than a leader without such power. Wolman (1956) found a strong relationship between a leader’s power and the leader’s acceptability to members as seen by observers and in peer ratings. Ziller (1955) showed that group opinion was influenced more strongly by a high-power figure than by a low-power figure. Similarly, Levinger (1959) found that the perceived relative power of a group member correlated .55 with number of attempts to influence others, .51 with range of assertiveness, and .48 with degree of assertiveness. A change in attempts to influence tended to change along with a change in perceived power as the group continued with the problem. Dahl, March, and Nastair (1957) observed the power of some U.S. senators whose announced roll-call votes influenced other senators in making their final decisions. According to Krupp (1986), the power of school principals provided them with the influence “to develop staff to high levels of productivity and enthusiasm” (p. 100).
Michener and Burt (1975a) examined determinants of leaders’ success in inducing compliance. They found that compliance was greater when leaders had reward power (i.e., justified their demands as being good for the group), had coercive power (i.e., could punish those who did not do as they had asked), and had a legitimate right to make demands on subordinates. However, neither the success nor failure of the groups, nor the approval of the leaders by their subordinates affected the leader’s ability to be influential. Consistent with this, Gainson (1968) suggested that leaders would shift toward coercion of subordinates if they felt they lacked the subordinates’ approval but had legitimate authority to ask for compliance, although such coercive power is often counterproductive.
Power can be counterproductive, undermining one’s efforts to be a leader, because it can be threatening, particularly in unstructured situations. A. R. Cohen (1953, 1959) found that in the face of powerful leaders, followers felt more threatened in unstructured than in structured groups. Followers whose self-esteem was low also tended to feel more threatened in the presence of power, and this feeling was intensified in unstructured situations.
McNeese-Smith (1999) conducted a study of 19 nurse managers, 221 staff nurses, and 299 patients. The managers’ motivation for power was negatively correlated with their leadership behaviors, as was the staff nurses’ job satisfaction. Nevertheless, the nurse managers’ power motivation contributed positively to patients’ satisfaction. According to a study by Kirchmeyer (1990) of 225 managers of both sexes, women, but not men, who were higher in need for power were more active in office politics. Men who played office politics were high self-monitors who felt that they worked in a difficult world. Matusak (1997) suggested that in many organizations, people strive to posses and guard their power, which they see as a valuable personal possession limited to a few. But in a trusting workplace, sharing power is seen positively and is unlimited, as will be discussed in Chapter 12 when empowering others is considered.
Etzioni (1961) conceived of power as physical, material, or symbolic. He defined these three forms of power as follows, giving the bases of each:
Coercive power rests on the application or … threat of … physical sanctions … ; generation of frustration, … or controlling through force the satisfaction of needs. … Remunerative power is based on control over material resources and rewards. … Normative power rests on the allocation and manipulation of symbolic rewards and deprivations.
Parsons (1951) held that normative power could be based on esteem, prestige, and ritualistic symbols or on acceptance. Etzioni also pointed out that it would be wasteful for a leader to emphasize remunerative power if the followers were already committed to the leader’s choice of action, since only normative symbolic rewards would be appropriate for such followers. Highly alienated followers would be inclined to disobey, despite material sanctions. Followers would be more likely to consider normative power legitimate and least likely to accept coercive power as legitimate.
French and Raven (1959) identified five kinds of power that quickly became popular among investigators as a way to type variations among the bases of power. Their five bases were as follows:
1. Expert power is based on B’s perception of A’s competence.
2. Referent power is based on B’s liking or identification with A.
3. Reward power depends on A’s ability to provide rewards for B.
4. Coercive power is based on B’s perception that A can impose penalties for noncompliance.
5. Legitimate power is based on the internalization of common norms or values.
Measurement. Several earlier survey questionnaires (e.g., Hinkin & Schriesheim, 1989) preceded the Interpersonal Power Inventory (Raven, Schwarzwald, & Koslowsky, 1998). Rahim (1988) developed a “leader power” questionnaire from the answers of 1,256 respondents. The final inventory contained five factorially independent scales to measure each of the bases of power. Subsequent validations were completed with a national random sample of 476 executives and a sample of 297 employed college students.
Problems with the Five-Base Model. The French and Raven model had a number of problems. First, the five bases are not conceptually distinct. For example, expertise is valued highly, so that sources of personal power in bases of expert power and referent power are likely to be correlated empirically, that is, lodged in the same people. In the same way, the position holder with the power to reward is also likely to have the power to punish. The position will give some degree of le-gitimacy as well. By definition, formal hierarchies are structures of legitimate, reward, and coercive power relationships.
Further conceptual difficulties were noted by Patchen (1974), who suggested that French and Raven’s classification was inadequate because the various bases of power were not conceptually parallel. Thus reward power and coercive power were defined in terms of resources that the influencer could apply. Referent and legitimate power were defined in terms of the characteristics and motives of the target person. Expert power depended on the characteristics of the influencers and on the information resources they personally possessed.
It was not surprising, therefore, to find that the five bases of power are empirically correlated. Rahim (1986) obtained, in a national random sample of 477 executives, a correlation of .58 between the expert and referent power the executives attributed to their superiors. The superiors’ control of rewards and coercive punishments correlated similarly with their legitimate power according to their subordinates. For 280 employees in 45 organizations rating the power of their 118 supervisors, At water and Yammarino (1996) reported correlations ranging from .33 to .55 among referent, expert, legitimate, and reward power. Coercive power correlated .51 with legitimate power and .41 with reward power.
Student (1968) regarded referent power and expert power as individually derived but thought that reward power, legitimate power, and coercive power were organizationally derived. He found that supervisors’ scores on incremental influence, referent power, and expert power were positively and significantly related to the quality of the group’s performance and to a reduction in costs. But average earnings declined with the reward power of the supervisor, and the maintenance costs of the group rose with coercion.
Podsakoff and Schriesheim’s (1985) review found expert and referent power positively correlated with each other as well as with employee performance, attendance, retention, and satisfaction with supervision. Legitimate, reward, and coercive power were intercorrelated and less effective. The empirical correlations between the bases of power make sense because power begets power. Power from one base can generate power in another. For example, individuals whose power is based on the legitimacy of their positions can acquire additional power by controlling the rewards that accompany the legitimacy of the positions. In the same way, people tend to defer to those they perceive to be experts. Perceived expertness, in turn, tends to legitimate the leadership role (Goodstadt & Kipnis, 1970). Although we may be able to sort out conceptually the different bases of power, in nature they are likely to be intertwined.
Filley and Grimes (1967) emerged with a finer set of distinctions. They studied 44 full-time professional employees in a nonprofit organization who reported to a director and to an associate director. The employees were interviewed about hypothetical incidents that might require them to seek a decision from the director or the associate director. The respondents were asked to whom they should go for a decision, to whom they would like to go, and to whom they would in fact go, and why. Answers to the “why” questions were classified according to the bases of power to which the organizational members appear to have responded. Twelve different bases were discerned. For example, instead of one category for legitimate power, several emerged. Legitimate power could be based on formal authority, on responsibility, on control of resources, on bureaucratic rules, or on traditional rules. Raven (1993) added equity, reciprocity, and dependence as bases for power. If A could control the equitable distribution of rewards to B, then A had power above and beyond reward power. If A could exchange one good for another with B, then A had the power of reciprocation. And if B depended on A, then A had power over B. An old adage is “Knowledge is power,” and in the information age it is not surprising to suggest that if A has knowledge personally or from his or her position that is desirable or useful to B, then A has informational power over B. Ordinarily, A is a leader and B is a follower, but the roles can be reversed.
Yukl and Falbe (1991) looked at the differences between lateral and downward power. In interviews with the peers of 49 middle and lower managers, persuasiveness was an important additional power. Persuasiveness, legitimate power, and expert power were the most important reasons mentioned by employees for their compliance with supervisor and peer requests. Rahim (1989) showed that expert and referent power were associated with compliance and satisfaction, while legitimate power was positively correlated with compliance but negatively correlated with satisfaction.
Hinken and Schriesheim (1989) made sharper distinctions in the French and Raven model:
Expert power is the ability to administer to another person information, knowledge, or expertise.
Referent power is the ability to administer to another person feelings of personal acceptance or approval.
Reward power is the ability to administer to another person things he or she desires or to remove or decrease things he or she does not desire.
Coercive power is the ability to administer to another person things he or she does not desire or to remove or decrease things he or she does desire.
Legitimate power is the ability to administer to another person feelings of obligation or responsibility.
Table 11.1 is a list of statements that emerged in factor analyses of the extent to which respondents agreed that the statements described their supervisors. The three samples of over 500 respondents were part-time employees enrolled in undergraduate business courses, full-time employees at psychiatric hospitals, and full-time employees taking MBA courses.
Table 11.1 Final Scale Items Responding to the Statement “My Supervisor Can …”
1. Expert Power Give me good technical suggestions. Share with me his/her considerable experience and/or training. Provide me with sound job-related advice. Provide me with needed technical knowledge. |
2. Referent Power Make me feel valued. Make me feel like he/she approves of me. Make me feel personally accepted. Make me feel important. |
3. Reward Power Increase my pay level. Influence my getting a pay raise. Provide me with special benefits. Influence my getting a promotion. |
4. Coercive Power Give me undersirable job assignments. Make my work difficult for me. Make things unpleasant here. Make being at work distasteful. |
5. Legitimate Power Make me feel that I have commitments to meet. Make me feel like I should satisfy my job requirements. Give me the feeling I have responsibilities to fulfill. Make me recognize that I have tasks to accomplish. |
SOURCE: Adapted from Hinkin and Schriesheim (1989, table 1).
In 1597, Sir Francis Bacon declared that “knowledge is power.” Expert power, perhaps unseen, can lie behind effective leadership. Gardner (1986a, p. 12) quoted Lyndon Johnson as saying, “When the press talks about my successes as Senate majority leader they always emphasize my capacity to persuade, to wheel and deal. Hardly anyone ever mentions that I usually had more and better information than my colleagues.”
Expert power may be manifest in information, knowledge, and wisdom; in good decisions; in sound judgment; and in accurate perceptions of reality (Watts, 1986). An item that is highly loaded (.78) on a factor of expert power is, “Has considerable professional experience to draw from in helping me do my work” (Rahim, 1986). In comparison with other bases, expert power appears to be most acceptable to and most effective with followers. It most readily gains their compliance and is least likely to provoke their resistance (Podsakoff & Schriesheim, 1985a).
The power of revolutionary or reform leaders often begins with their perceived power as experts, which they use to define prevailing problems and to develop innovative solutions. Followers are persuaded that the leaders have the right answers to their problems and are organized to provide support (Gjestland, 1982).
In the physician-patient relationship, the physician wields expert power and the patient plays a distinctly subordinate role, particularly if the patient is also lower in socioeconomic status (Fisher, 1982). But whether the patient complies with the physician’s expert advice will depend on what that advice entails as well as the physician’s referent power. Members of a group who have relevant information about a task will attempt to lead the group (Hemphill, Pepinsky, Shevitz, et al., 1954). Such attempts are likely to be successful if the members are perceived to be expert, and effective if they really have the expertise. Groups tend to defer to the actual and the perceived expert. They are likely to be persuaded by the perceived expert, to accept the expert’s opinion both publicly and privately (Bass, 1960).
H. T. Moore (1921) obtained experimental support for the idea that we readily accept the influence of those whom we accept as experts. Moore observed that students shifted their judgments (about linguistic, ethical, and musical matters) toward what the experimenter led them to believe was the opinion of experts. When Mausner (1953) introduced one student as an art student and another as an art expert, the group’s opinion was more strongly influenced by the art expert than by the art student. Knight and Weiss (1980) arranged for leaders of task groups to be seen as “expert” or “nonexpert” on the basis of how they were chosen. Those who were chosen as experts were better able to influence the group than were those who were chosen as nonexperts. The selection of leaders from inside or from outside the group made no difference.
In another experiment, Mausner (1954b) demonstrated that subjects tended to agree more often with a partner whom they observed to succeed than with one whom they observed to fail. Luchins and Luchins (1961) demonstrated that an expert’s opinion was more influential than the majority opinion of the group in determining the group’s response to a judgmental task. Torrance (1952) and M. A. Levi (1954) both obtained greater improvement on survival problems when groups were fed back expert information after an initial training period.
Evan and Zelditch (1961) studied experimental groups with supervisors who differed in their knowledge of the task. Differences in the supervisors’ knowledge did not affect the group’s productivity; however, group members exhibited more covert disobedience and resistance to the least informed supervisor than to the moderately or well-informed supervisor. This effect was attributed to changes in the followers’ attitudes toward the right of the poorly informed supervisor to occupy a position of leadership. Similar results were reported by G. M. Mahoney (1953), who found that workers were much more satisfied with a wage-incentive system when they thought the supervisor did a good job of explaining the reasons for changes in the system. A survey of more than 200 graduate students found that faculty members with a large amount of expert power were more likely to be acceptable to the students as thesis advisers, coinvestigators in research, and coauthors of conference papers. This was in contrast to potential faculty advisers who were perceived as coercive (Aguinis, Nesler, Quigley, et al., 1996).
Caveats. The impact of expert power has its limits, as was noted in Chapter 4: would-be leaders must be able, but not too much more able than their followers. Many examples can be cited of experiments in which experts failed to be influential. C. B. Smith (1984) observed that apparent technical and administrative expertise in a contrived short-term employment situation did not increase the experimental subjects’ output as coders or their compliance with directives. Collaros and Anderson (1969) studied groups with one expert, all experts, and no experts. Groups in which all members were told that they were experts were more inhibited in their performance.
Although individuals can become more influential by acquiring more knowledge and expertise, the technological revolution in the spread of knowledge, and the ease with which information can be transmitted, can quickly alter who is expert about the available information and thus alter the resulting patterns of influence. Under some circumstances, then, a clerk with access to a computer program may become more expert and influential than a PhD without access to the program. Cleveland (1985) noted that information is expandable, compressible, substitutable, transportable, diffusable, and shareable. It is not necessarily a scarce resource. Expert power is more fluid than fixed in a position or a person. It can be found in a computerized database, a computer software program, or in a search engine such as Google. Custom-made expert programs can substitute for human expertise. Increasingly, research focuses on information power, which may differ from expert power as such.
Some further qualifications are in order. The stronger status-leadership relationship can override expert power. Torrance (1955a) concluded that the high status of a team member contributed more to influencing other members to accept the correct answer to a problem than did knowing the correct answer. Paradoxically, two supposed experts may generate less confidence than one alone. Torrance and Aliotti (1965) studied groups involved in information seeking with one or two randomly selected students in the role of expert. They found that groups with two “experts” obtained more accurate information than did those with one “expert,” but the groups with two “experts” were less certain of their judgment.
Liked, respected, and esteemed leaders have referent power. Followers want to identify with leaders who have referent power and be accepted by them. The evidence is primarily indirect about the extent to which referent power contributes to influence and leadership. Like expert power, referent power is correlated with the ten-dency to use rational tactics (Hinkin & Schriesheim, 1990). Referent power has become more important with the downward shift of power to lower organizational levels and the increasing importance of personal values and convictions in the influence process (Behr, 1998). Being esteemed and valued by followers is highly related to leaders’ referent power. Yukl, Kim, and Falbe (1996) asked 195 full-time-employed MBA night students to describe three critical incidents in their organizations. In these incidents, leaders who had more referent power were also likely to generate more commitment. They were less likely to use pressure tactics. Hinkin and Schriesheim (1990) reported similar findings, as well as a substantial direct relation-ship between referent (and expert) power and the use of rationality as an influence tactic.
We need to consider indirect evidence of referent power due to ingratiation and the desire for acceptance by followers. For much of this indirect evidence, referent power is assumed from friendship or popularity. Thus Rahim (1986) found factor loadings for referent power of .81 and .85 for such items as, “I like the personal qualities of my superior” and “My superior is the type of person I enjoy working with.” Studies of referent power need to capture its essence in the desire of the less powerful to identify with those who have referent power (Podsakoff & Schriesheim, 1985a). For example, referent power is clearly seen when a charismatic relationship exists between the leader and the follower, since what binds the follower to the leader is the desire to identify with the leader.
Esteemed individuals are more likely to emerge as leaders. In 17 studies of executives in various organizations, esteem—as estimated by merit ratings received from superiors—was highly related to “real-life” success as a leader (Bass, 1960). The same has been found in controlled experimental settings, but with qualifications based on selected experimental manipulations of referent power. French and Snyder (1959) found that more highly accepted leaders in experimental groups attempted to exercise more influence, and had more effective groups, than did those who were less well accepted. Followers also attempted to use more influence when they were accepted by their leader. In a review, Podsakoff and Schriesheim (1985a) found that the use of referent power by leaders usually contributed to better performance, greater satisfaction, greater role clarity, and fewer excused absences among their subordinates.
Acceptance, Popularity, and Power. Zander (1953) and Zander and Cohen (1955) introduced two strangers to groups. The one introduced as high in prestige felt better accepted and more at ease than the one introduced as in a low-prestige role. Hurwitz, Zander, and Hymovitch (1953) combined and recombined members of six-person discussion groups. Each individual rated the others on the degree of liking and perceived power and participation. Members with a high degree of power were better liked and participated more often than those with a low degree of power.
Ingratiation. As mentioned earlier, Bass, Wurster, and Alcock (1961) demonstrated that we want to be esteemed by those we esteem. We are more concerned with being liked and accepted by those we respect and accept—those with referent power. This concern, in turn, may lead to ingratiation—the striving by followers to be valued and rewarded by those they esteem or see as more powerful. A line of investigation by Jones and Jones (1964) indicated that followers who want to be liked by high-power figures employ subtle ingratiation tactics, such as flattery, in an effort to gain acceptance. Jones, Gergen, Gumpert, et al. (1965) showed that participants in an experiment who faced the prospect of poor task performance attempted to ingratiate themselves with the experimental supervisor by presenting themselves as strong and competent, but only if the supervisor was open to influence. L. Wheeler (1964) demonstrated that low-power subjects remembered more autobiographical statements made by the high-power figure. Wheeler did not find that participants who were highly dependent on their task leader sought information as a means of ingratiation or to increase their own power. But A. R. Cohen (1958) demonstrated that low-status members who could increase their status in the group tended to communicate in friendly, ingratiating ways. These tactics were designed to protect and enhance their relations with those who controlled the upward-mobility process. But those members with little perceived opportunity to increase their status made relatively few such attempts. There may be a rational payoff in ingratiating behavior by followers. Thus Kipnis and Vanderveer (1971) observed that leaders tend to reward their ingratiating followers.
Referent power can become coercive. Moreno (1934/ 1953) postulated that if A highly valued B, B could injure A. And Bass (1960, p. 289) described how referent power could be subverted:
[If] you can give me affection, self-esteem through association with you and vicarious satisfaction by identification … although I may not privately accept what you say, I will publicly agree with you so that you will grant me what I want from you … (And such referent power may be lost as a consequence) … because if I must continue to inhibit my own opinions … I will begin to value you less, to dislike you and eventually reject you.
Reward power implies an ability to facilitate the attainment of desired outcomes by others. An item highly loaded on a factor of reward power is, “Can recommend a promotion for me if my performance is consistently above average” (Rahim, 1986). Marak (1964) studied some groups who were rewarded for a correct decision and others who were not. The results indicated that the ability to provide rewards was related to leadership, as measured by sociometric, interaction, and influence scores. The more valuable reward a member could provide, the more closely this ability was related to leadership. Evidence for the emergence of a leadership structure was suggested by the finding that attempted leadership, actual influence, and rewards for initiating leadership increased as the sessions progressed. Similarly, Herold (1977) showed that the behavior of subordinates in 32 trios depended on the manipulation of monetary rewards by the trio leaders. In a leader-subordinate experimental simulation, Dustin and Davis (1967) observed that when given a choice, leaders used monetary rewards twice as much as they used praise. Kipnis (1972) also found that economic incentives were favored over other ways of improving subordinates’ performance. As before, Hinton and Barrow (1975) found that when subordinates performed at high levels, supervisors tended to make more use of economic reinforcements than of praise. However, when subordinates performed poorly, leaders tended to make more use of reproof.
Many of the critical behaviors that separate successful from unsuccessful noncommissioned officers in situational tests were found by Flanagan, Levy, et al. (1952) to be due to the differential reinforcements provided by these officers. Successful officers more often encouraged their men to follow rules and regulations, gave pep talks when the men were tired, and constantly checked the behavior of their men. However, S. Kerr (1975) noted that in their desire to achieve one kind of behavior, leaders as well as organizations sometimes unintentionally reward another kind.
Greene (1976a, 1976b) conducted the first longitudinal study of the effects of rewarding behavior by leaders and concluded that it could result in improved performance by the subordinates. Sims (1977), Sims and Szilagyi (1978), and Szilagyi (1980b) also found a causal relationship between such reward behavior and follower performance in a series of longitudinal studies.
Justis (1975) examined effective leadership as a function of the extent to which followers’ rewards depended on the leader’s competence and performance. The leader’s effectiveness and influence were greater the more the leader was seen to be competent and the more the followers’ rewards depended on the leader’s performance. Expert power interacted with reward power.3
Interdependence Effects. Berkowitz (1957b) experimented with pairs of participants, one or both of whom could earn a reward. The perception of interdependence increased the members’ motivation when both were eligible for valued rewards. Participants were also motivated to work toward a partner’s reward even when they were not eligible. Berkowitz and Daniels (1963) demonstrated that participants worked harder for a “supervisor” whose success was dependent on their performance than when no such interdependence existed. Berkowitz and Connor (1966) varied both success and dependence for pairs who could win rewards. Participants who experienced failure expressed stronger dislike for their partner the greater their feeling of responsibility toward the partner. Successful participants, however, worked harder for their dependent partner than did the control participants.
The leader may also have to share the power to reward with other authorities, especially in matrix organizations where a subordinate reports to a functional leader and a project leader. In such a case, Hinton and Barrow (1975) found that leaders were more likely to use their reward power if they had to share its use than if they operated alone.
Distributing Rewards. A norm that is often common to students and workers in the same small-group settings is “Share and share alike.” Thus Morgan and Sawyer (1967) found that pairs of boys, when permitted to earn equal or unequal rewards, preferred equal rewards whether their partners were friends or strangers. But their perception of the other’s expectation played an important part in determining their preference. Even ingratiating subordinates who disparaged the competence of their student supervisors failed to modify the allocation of rewards by the supervisor (Fodor, 1974). And when Shriver (1952) gave leaders of discussion groups checks of various amounts with which they were required to reward the contributions of other members, some leaders solved the problem by drawing lots; others, after delay and emotional upset, passed out the checks quickly and departed. Shriver interpreted these results as indicating that there are limits to the reward power that an emergent leader can exercise with comfort. Nevertheless, W. P. Smith (1967b) found that participants with a highly valued outcome liked their partners less than those with less highly valued outcomes. Those with little power valued their outcome more highly and used their power more positively than did those with much power.
Yet considerations of equity and loyalty do have an impact on the situation. Thibaut and Faucheux (1965) varied equity and loyalty to one’s group in bargaining for individual gain. Partners who were told that they had greater power tended to use it; other partners tended to acquiesce. Lower-power members appealed for “a fair share” when the high-power members could manipulate rewards, but they appealed to “loyalty” when rewards were manipulated by an outsider. Thibaut and Cruder (1969) found that participants tended to form contractual agreements when they discovered that an agreement restricted the power of each to prevent their joint attainment of maximum outcomes. P. Murdoch (1967) observed that members tended to develop contractual norms when the divider of rewards was presented as being likely to withdraw from the relationship. Butler and Miller (1965) and McMartin (1970) determined that participants tended to distribute rewards in proportion to the difference in the average rewards received from others. Swingle (1970a) demonstrated that cooperative participants were exploited more when they were powerful than when they were weak, even when exploitation resulted in reduced rewards for the exploiter.
Power affects promises to cooperate for mutual benefit. Tedeschi, Lindskold, Horai, et al. (1969) studied reactions to participants who repeatedly promised to cooperate but who varied in credibility (their degree of cooperation following their promises). Powerful teammates ignored the promises and failed to cooperate. Participants equal in power were most cooperative with the willing member. Participants weak in power became more exploitative as the credibility of promises increased. L. Solomon (1960) demonstrated that unconditional offers of cooperation by one member equal in power to a partner was exploited by the partner who became less liked. However, if the offers to cooperate were conditional, more cooperation and more liking of the partner were elicited. If the partners were unequal in power, the results were reversed.
Antecedents to Usage and Effects. Hinton and Barrow (1976) showed that reward behavior was most frequently used by responsible, confident, and enthusiastic leaders. Barrow (1976) and Herold (1977) found that leaders rewarded good performers and were more punitive toward poor performers. Hunt and Schuler (1976), Oldham (1976), and Sims (1980) obtained similar results. But Greenberg and Leventhal (1976) reported that leaders offered financial bonuses to poor performers to motivate them when that was the only motivator available.
Bennis, Berkowitz, Affinito, and Malone (1958) studied reward power in hospitals. They found that supervisors at hospitals where the rewards that were given were congruent with those hoped for by the subordinates exercised more influence than did supervisors at hospitals where the rewards were not congruent with the subordinates’ hopes. Furthermore, the hospitals that gave congruent rewards were more effective. But supervisors exhibited little awareness of the subordinates’ preferences for rewards. A participative style of leadership (see Chapter 18) was reported by Singh-Gupta (1997) among Indian managers correlated with their use of expert and reward power.
To appreciate fully the effects of reward power, one must consider how it is used. On the one hand, many studies have demonstrated the utility of rewarding by supervisors in exchange for compliance by subordinates. Supervisors’ recommendations for rewards that are contingent on the subordinates’ performance have been widely found to contribute to productivity and to effective operations. However, such rewards may be resented by subordinates and may actually be coercive. In such cases, compliance will be public only, not private, especially when subordinates see the supervisors’ use of rewards as capricious, arbitrary, and unfair, rather than reasonable, predictable, and fair.
Epictetus declared:
… no one is afraid of Caesar himself, but he is afraid of death, loss of property, prison, and disenfranchisement. Nor does anyone love Caesar himself unless in some way Caesar is a person of great merit; but we love wealth, a tribuneship, a praetorship, a consulship. When we love and hate and fear these things, it needs must be that those who control them are masters over us. … That is how at a meeting of the Senate a man does not say what he thinks, while within his breast his judgment shouts loudly. Starr, 1954, p. 144
The leader who uses coercive power controls the granting or denying of valued rewards or feared penalties; subordinates’ private opinions and feelings remain hidden, but there is pressure on them to express publicly what they really feel. In hierarchical settings, coercion is manifest when the subordinate “holds in abeyance his own critical faculties for choosing between alternatives and uses the formal criterion of the receipt of a command or signal as his basis for choice” (Simon, 1947, p. 126). Extremes in coercive power were seen in Shaka, the Zulu king, who was known to have summarily executed his courtiers for a breach of etiquette, a smile at the wrong time, disagreeing with him over a minor point, or one of many other slight causes of his displeasure (Morris, 1966). As Matusak (1997) noted, “Power without caring, commitment, and empathy can become a selfish tyranny.”
Tannenbaum (1950) listed ways in which executives may use their power to restrict or inhibit subordinates’ behavior. The executives may arbitrarily identify the organization’s goals, set up criteria for evaluating alternative paths to the goals, rule out alternatives, limit the general activities in which subordinates are permitted to engage, identify the positions with control and power, give or withhold information, and set deadlines to be met to avoid punishment or earn rewards.
Among Indian managers (as well as elsewhere), authoritarian leadership is associated with the use of coercive power (Singh-Gupta, 1997). Coercive power implies the ability to impose penalties for noncompliance. Rahim (1986) found that endorsement of the statement “My superior can fire me if I neglect my duties” correlated .82 with a factor of coercive power. French and Raven (1959) reported that conformity by followers (public acceptance but private rejection) was a direct function of earlier threats of punishment for noncompliance. Although coercion most commonly involves punishment or its threat, more subtle uses of power to coerce may involve promising rewards for compliance. One will comply publicly but perhaps not privately as a consequence of such promises and one’s concern about failing to obtain promised rewards in the absence of compliance. Thus A has power over B if A can control whether B is rewarded or punished and B seeks rewards and wants to avoid punishment. A is coercive if B behaves publicly according to A’s demands, although B privately rejects A (Bass, 1960). A clear demonstration of coercion that results from power was reported by French, Morrison, and Levinger (1960). In laboratory assignments, participants serving as subordinates exhibited a much greater discrepancy between their public and private reactions when supervisors could assess monetary fines than when no fines were established.
Militant leaders of social and political movements apply coercive pressure on officials who are vulnerable to their attacks, such as “high-minded” university presidents, elected politicians, or executives whose business may suffer. These militant leaders gain visibility and mobilize their supporters to support their coercive efforts, which are directed at changing the actions of their targets before changing the targets’ attitudes. Such leaders harass, threaten, cajole, disrupt, provoke, and intimidate in speeches, dress, manners, gestures, slogans, rituals, and violent confrontations (Simons, 1970).
Usage. The use of coercive power is less popular with leaders and subordinates alike. Leaders tend to use this form of power to deal with unacceptable performance by a subordinate when they do not have the power to reward the subordinate for acceptable performance (Kipnis, 1976) or do not have other bases of power. But even in this instance, there is reluctance to use coercive power. McFillen and New (1979) demonstrated that leaders were less willing to exercise monetary sanctions for poor performance when they had only the power to punish. They were more willing to use monetary rewards for good performance when they had only the power to reward their subordinates. Leaders were more inclined to use their coercive power when they were under pressure to maintain high-productivity schedules and had lost their power to reward good performance (Greene & Podsakoff, 1981). Nonetheless, Simon (1947) argued that the use of coercive power is hard to avoid in any formal hierarchy, for five reasons: (1) subordinates develop expectations for obedience to symbols of higher status; (2) the superior can satisfy the subordinates’ personal need for security from a substitute father or mother; (3) the subordinates may share the same goals as the superior and perceive that blind obedience provides a means to obtain the goals; (4) the subordinates are freed from the responsibility for making difficult decisions, and the superior bears the burden of these decisions for them; and (5) most simply, the superior may be able to reward or punish the subordinates materially.
Subordinates in a business organization are coerced by others besides their immediate superiors. They may have to accept publicly, yet reject privately, the dictates of buyers of the organization’s products; contractual agreements with labor unions or industrial cartels; and the demands of government, custom, and tradition (Bass, 1960). More often than not, coercion in a hierarchical setting is subtle. Subordinates may not even be aware that they are being coerced (Timasheff, 1938) and may repress any private feelings that are at odds with their public statements. For example, an executive may be reminded by the CEO of the costs in lost stock options if the executive quits. Or a young executive with new ideas may be coerced by a superior who keeps emphasizing the possibility of failure (Kets de Vries, 1980).
Although it may not be very subtle, pure political influence can be coercive when it displaces the rational and legitimate use of power with Machiavellian deception, divisiveness, defensiveness, or emotional appeals. Even an offer of rewards for political favors can be seen as coercive and may require a strong countervailing conscience to resist. Power can be wielded by political manipulations, such as covertly denying, delaying, or distorting information that is sent to another member of the organization so that the latter’s choices are restricted (House, 1984).
Indirect Coercion. Observing vicariously how penalties are tied to the performance of one’s colleagues may be more effective than direct threats in maintaining performance in the short term. Schnake (1986) studied the effects of punishment on the attitudes and behavior of coworkers who observed a peer receiving punishment. Students hired for temporary clerical employment observed a coworker receiving a reduction in pay, a coworker receiving a threat of a reduction in pay, or a coworker who was not penalized. Those who observed the coworker receiving a reduction in pay produced significantly more output than those who observed the coworker being threatened with a reduction in pay or not being penalized. Job satisfaction was not affected, and the effects held for at least a week.
“It’s Not What You Say, But the Way You Say It.” Compliance can actually be due to coercion even if the person who complies does not feel coerced, as a consequence of the language used in the attempt at influence. Drake and Moberg (1986) detailed how both public and private compliance can result from the language used by A to obtain B’s compliance when ordinarily B would see A’s request as exploitative and coercive. For instance, A can use sedating language, which suppresses B’s tendency to analyze whether B will gain or lose by complying. Or A can be indirect (“Something needs to be done about the trash”), rather than direct (“Take out the trash”); in this case, A becomes the observer, not the leader. Also, A’s observation of the existence of a problem and related hints, prompts, and teases can substitute for a direct order. A can first determine B’s availability and willingness to comply before assigning new responsibilities to B. (But if A behaves this way routinely, B may interpret A’s effort as a characteristically annoying ploy and hence may lose trust in A.)
Giving orders or making requests without explanation is likely to produce less compliance and more sense of being coerced than giving logical reasons. However, more compliance will occur and less coerciveness will be felt even if the reasons do not really make any sense. Such reasons can be substituted because cognitive processing is limited when decisions are made to follow orders in many seemingly legitimate situations. For example, Langer, Blank, and Chanowitz (1978) showed that personnel who were using a copying machine would allow a usurper to take over the machine for an irrelevant reason (“I have to make copies”) but they would comply less often if no reason was given.
Power is legitimated in the language used. Hence, A increases power over B by using rich, rather than redundant, vocabulary and by expressing less uncertainty in appeals (Berger & Braduc, 1982).
Another approach to offset the potential coerciveness of a direct order is to precede it with disclaimers. As quoted by Drake and Moberg (1986, p. 578), Hewitt and Stokes (1975) outlined five forms of disclaimers:
Hedging (I’m not really committed): “I haven’t given this much thought but …”
Credentialing (I’m not prejudiced): “Some of my best friends are …”
Sin licensing (I’m not a rule breaker): “I know this is against the rules but …”
Cognitive disclaimers (I’m not confused): “This may seem strange but …”
Appeals to suspend judgment (Don’t get offended): “Don’t get me wrong but …”
Excuses and justifications may also be used to palliate feelings of exploitation and to increase compliance. Compliance can be increased by avoiding messages of powerlessness. O’Barr (1982) reported that more convincing witnesses in court trials avoided using power-less language—intensifiers (e.g., “very”) and hedges (e.g., “sort of”).
Conditions That Increase the Use of Coercive Power. Coercive power dominated supervisory-subordinate relations in a factory where workers were marginal, handicapped, without skills, and unable to get jobs elsewhere (Goode & Fowler, 1949). Kipnis and Cosentino (1969) compared corrective actions used by military and industrial executives. Industrial supervisors tended to reprimand and to transfer offending employees, whereas military supervisors more often used instruction, reassignment, extra work, and reduced privileges as corrective measures.
Boise (1965) studied attitudes of supervisors in a city’s police, street, and water departments toward correcting subordinates’ peformance. All but authoritarian supervisors (those with high F scores) and police supervisors agreed with the concept of a uniform penalty for unacceptable performance. Coercive power is most likely to be used to deal with noncompliance. Katz, Maccoby, Gurin, and Floor (1951) found that the foremen of low-producing railroad section gangs were more punitive. A longitudinal study by Greene (1976b) concluded that supervisors are punitive in reaction to their subordinates’ poor performance. Szilagyi (1980b) established that higher levels of punishment tended to follow higher levels of absenteeism by employees. Bankart and Lanzetta (1970), Barrow (1976), and Hinton and Barrow (1975) also reported the tendency of supervisors to become coercive as a consequence of subordinates’ inadequate performance. Situational variables, such as difficulty of the task and prospects for failure, increase a leader’s tendency to employ coercion (Michener, Fleishman, Elliot, et al., 1976). That is, leaders who face failure will tend to become more coercive (Kipnis, 1976). T. R. Mitchell (1979) further showed that more coercion was practiced if the supervisor attributed poor performance to a lack of motivation rather than to a lack of ability, or to policy changes or bad luck.
Goodstadt and Kipnis (1970) studied student work groups under different types of supervisory power. Student supervisors tended to use coercive power to solve disciplinary problems, but they used expert power to solve problems arising from ineptness. Sims (1980) concluded from a review of both longitudinal field studies and laboratory studies that punishment tends to be more a result than a cause of employees’ behavior. More specifically, managers tend to increase punishment in response to the poor performance of employees. Illustrating the use of punishment in reaction to other events, including the loss of other bases of power, Greene and Podsakoff (1979) found that after a contingent (positive reinforcement) pay plan at two paper plants had been abandoned, supervisors began to use rewards less and punishment more. Fodor (1976) found that leaders were more authoritarian and coercive with disparaging and disruptive subordinates. The leaders rated such subordinates lower and gave them less pay than they gave to other subordinates. In the same study, Fodor noted that supervisors who were subjected to group stress also tended to become more coercive and less rewarding.4
Support for Coercive Power. Legitimacy coupled with coercion will increase public and private acceptance of coercive demands. In a study of experimental groups, Levinger, Morrison, and French (1957) found that the threat of punisment induced conforming behavior in group members before punishment but not after punishment. Most to the point was the fact that the group members’ perceptions of the legitimacy of the punishment reduced their resistance to conform. Particularly important is the support of the status structure. When Iverson (1964) presented taped speeches by people who were identified as high or low in status, those who listened to speakers who made punitive remarks formed more favorable impressions of the personalities of the high-status speakers than they did of the low-status speakers. The coercive demands of an attractive esteemed leader are more likely to be accepted, although continued use of coercive power will reduce such attractiveness and esteem. Another support for ready obedience to coercive power is the immaturity of participants. Kipnis and Wagner (1967) found that immature participants performed better than did mature participants under a leader whose decisions to administer punishment were backed up by superior authority. Brass and Oldham (1976) examined the reactions of first-line supervisors to an “in-basket” simulation.5 Overall, the supervisors punished more often than they rewarded their subordinates. The more active supervisors, those who used both more reward and more coercive powers, were rated as more effective.
Leaders use coercion more easily when they rationalize that they are only following orders from higher authority without questioning the legitimacy of those orders. They are more coercive if they attribute the coercion to their position, rather than to themselves as persons, and if they can maintain sociopsychological distance by thinking of those they are coercing as enemies, inferiors, and troublemakers. Adolf Hitler made coercion acceptable to the Germans by consolidating the attention of the German people as the “master race” against a single adversary and then lumping different opponents together in the single category of those who were antithetical to the master race (Paige, 1977).
Resistance to Coercive Power. When Milgram (1965b) instructed students to activate a device that supposedly administered electric shock to subjects who made errors in a learning experiment, the students were willing to administer “dangerous” degrees of shock on command. The students had a strong tendency to obey authority and to accept being coerced by it. But when students who were administering the electric shock were permitted to observe two other individuals who refused to do so, the number of refusals was significantly increased. Stotland (1959) found that when participants were required to work with a domineering supervisor, they tended to identify with the supervisor. However, when they were able to interact with a peer from another group, they would exert more independence from the power figure, show more hostility toward him, and exhibit greater motivation to reach the goal despite the supervisor’s hindrance. French, Morrison, and Levinger (1960) observed that the greater the threat of punishment, the less the resistance to compliance, but the resistance increased after the threat disappeared.
A form of “group compensation” may arise, according to Stotland (1959), in resistance to attempts to be coerced if the opportunity to do so is made available. In Stotland’s study, individuals used models to design a city and were subject to the veto power of a “supervisor,” who interrupted their work twice. Some of the participants were given the opportunity to meet privately in the absence of the “supervisor;” others were not. Those who met privately became much more aggressive and hostile toward the supervisor, whereas those who did not continued to accept the supervisor publicly. Private assembly permitted the formation of an informal organization to present a unified front against the “supervisor’s” continued attempts to coerce the participants. In sum, group members were better able to resist coercive power when they had an opportunity to interact with peers in the absence of the supervisor or to observe other members who disobeyed orders. Resistance would be higher if the orders were judged to be illegitimate and the leader was personally unattractive.
Resisting attempts by superiors to exercise their coercive power may be productive. In an experiment by Horwitz, Goldman, and Lee (1955), students were frustrated by the arbitrary refusal of their teacher to repeat instructions. However, they shifted to more favorable attitudes toward the teacher, were less annoyed, and regarded the teacher as less coercive when the teacher read their opinions and agreed to repeat the instructions. These more favorable attitudes occurred less often when the teacher was informed by a “higher authority” that the instructions should be repeated. The students solved problems and learned best when they could take direct action to “reform” the coercive teacher; problem solving and learning were somewhat less efficient when the students could only state their grievances, and were poorest when the students were led to believe that all the other students accepted the coercive teacher’s refusal to repeat the instructions.
Unintended Consequences of Coercion. Those with coercive power will readily fall back on it when necessary. They are tempted to exploit their power by demanding a greater share of available resources and rewards. The process is reinforced by colleagues’ development of mistrust of the coercive power holder and by the power holder’s counterreaction of suspiciousness about the colleagues’ intentions, which results in his or her greater reliance on coerciveness (Kipnis, 1976). Coercive power can turn a mutually benevolent relationship into one of mutual hostility (Deutsch & Krauss, 1960). The costs of depending on coercive power in formal organizations are well known. For instance, satisfaction with supervision is almost uniformly lower. Thus whereas the satisfaction of life insurance agents was related positively to their manager’s use of referent and expert power, it was related negatively to their perceptions of the degree of reward, coercive, and legitimate power that the manager exercised (Bachman, Smith, & Slesinger, 1966). Weschler, Kahane, and Tannenbaum (1952) observed that a research laboratory headed by a coercive leader was more productive but less satisfying to its members than a laboratory that was under less coercive leadership. Zander and Curtis (1962) indicated that participants in a coercive situation found each other less attractive, less well accepted, and less motivated than they were in a referent situation. Weiss and Fine (undated) found that participants who had experienced insult and failure were receptive to arguments to be hostile and punitive, whereas comparable participants who had satisfying experiences were more ready to accept communications asking them to be lenient. Hostility toward power figures who can coerce other members was also found by Stotland (1959). Participants may withdraw from the problems at hand to avoid being coerced. Riecken (1952) attributed such withdrawal to the frustrations of coercion. Oldham (1976) related punishment by leaders to evaluations of their subordinates’ effectiveness. Generally, punishment was correlated not with subordinates’ performance, but with their lack of “motivational effectiveness.” Punitive leadership was found by Sims and Szilagyi (1975) and Sims (1977) to be uncorrelated with the subordinate performance of professionals and technical personnel, but it had a significant inverse correlation with the satisfactory performance of administrative and service personnel as subordinates.
A line of investigation beginning with Janis and Fesh-bach (1953) generally supported the contention that strong threats are less productive than moderate or mild threats. In the prototype experiment by Janis and Fesh-bach, conditions of strong, mild, and minimal fear were included in propaganda about dental hygiene, and resistance to counterpropaganda was measured. The arousal of minimal fear seemed to provide the most persistent influence, whereas the arousal of strong fear resulted in the ignoring or avoiding of attempts to influence (a form of withdrawal) or the development of conflicting tendencies to minimize the importance of the threatening propaganda. On the basis of similar findings, Kelman (1953) argued that when coercion is too great, interfering responses and hostility toward the coercer made participants less prone to accept the coercer’s attempts to influence them. Yet the crushing of the prodemocracy movement at Tiananmen Square in 1989 in China illustrates that credible strong threats followed by application of overpowering force and punishment can be a powerful deterrent. Nevertheless, subjugation to continued coercion is likely to contribute to dysfunctional outcomes such as learned helplessness, feelings of powerlessness, alienation, and depression. When possible, it can generate verbal and interpersonal aggression and provoke sabotage in those who are being coerced (Spector, 1975). In a study of 216 nurses and nurses’ aides, Sheridan and Vredenburgh (1978a) found that the use of coercive power contributed to heightened tension.
Caveat. Most of the research on coercive power was completed before 1980. Since then, the use of coercion has probably declined or become more subtle, owing to societal enlightenment, threats of lawsuits, liberalization of education, and public condemnation of glaring examples of abusiveness. Coercion is still more acceptable in authoritarian societies than in western democracies. But terrorism and military force remain strong as efforts to use coercive power to influence.
Legitimate power is based on norms and expectations regarding behaviors that are appropriate in a given role or position. The legitimatization of a role is derived from such norms and expectations. Members are more likely to accept the position of the leader and his or her influence as legitimate when the leader holds attitudes that conform to the norms of the group or organization. Read (1974) argues that the legitimacy of leaders involves a complex set of attitudes toward them and their source of authority; the leaders’ actual behavior contributes to their continuing legitimacy. Bleda, Gitter, and d’Agostino (1977) observed that the satisfaction of enlisted personnel with army life was related more to the leadership of those they saw as the originators, rather than as the relayers, of their daily orders.
French and Raven (1959) suggested three sources of legitimate power: (1) cultural values that endow some persons with the right to exercise power; (2) occupancy of a position organized to confer authority; and (3) appointment or designation by a legitimizing agent. An item that Rahim (1986) found to be highly correlated (.74) with a factor of legitimate power was, “My superior has the right to expect me to carry out her (his) instructions.”
According to Hollander (1993), “influence and power flow from legitimacy” (p. 29). By legitimatizing leadership, followers affect their leaders’ influence and behavior and the performance of the group. The legitimacy of leaders depends fundamentally on their standing with their followers.
In goal-setting experiments, students and workers usually try to do what they are asked (Latham & Lee, 1986); they generally regard the requests from experimenters in laboratories and supervisors in work settings to be legitimate. The experiment is a “demand situation.” In the workplace, employees think it is legitimate for their supervisor to tell them what to do—that this is inherent in the employment contract (Locke, Latham, & Erez, 1987). The legitimacy of supervisors’ orders was significantly related to the subordinates’ intention to work hard for an assigned goal in Oldham’s experiment (1975).
A survey of 1,155 soldiers who had completed basic training to prepare them for action in the Korean War showed that the soldiers had a stronger “fighter spirit” to the degree that they perceived their leader had legitimacy (R. B. Smith, 1983). According to Michener and Burt (1975a, 1975b), studies of college students indicated that recognition of the authority of the leader’s office was more important to their compliance than was endorsement of the leader’s personal right to exercise power.
House (1984) provided a rationale for the fact that although people accept leadership based on legitimate power, they do not do so with any special enthusiasm or as readily as they accept leadership based on expert power. When influenced by legitimate direction, followers are being responsive to the source, not the content, of the attempted leadership (Wrong, 1980). They tend to attribute the consequences of their compliance to external factors, not to themselves (Litman-Adizes, Raven, & Fontaine, 1978). They are less personally involved and can take less credit for and less satisfaction from success due to their compliance. They do what is expected of them but not more than may occur in response to leadership based on referent or expert power.
Legitimacy Through Appointment, Election, or Emergence. How the legitimacy of a position is established makes a difference. Appointment or election to a position tends to legitimate the leadership role to a greater extent than does emergence in the role through interaction or capture of the role by force. Thus in an experiment with 82 groups of three to five male undergraduates, Burke (1971) found that the basis on which leadership was legitimated by election, emergence, or counterelection was more important to the role differentiation that occurred within the groups than whether the goal was or was not established by consensus and whether pay was distributed equally or differentially. Huertas and Powell (1986) found an increase of ingratiation and conforming statements among members if a leader was not appointed but was allowed to emerge.
Legitimacy was manipulated in an experiment by Goldman and Fraas (1965). Leaders of groups playing “twenty questions” were either elected, appointed by reason of their competence, or randomly appointed. Control groups had no formal leader. The groups’ performance, as measured by the time required and the number of questions needed to reach the solution, was poorer if the leaders were randomly appointed or if no leader was put in place than if leaders were legitimated by election or appointment because of their competence.
Hollander and Julian (1969) and Firestone, Lichtman, and Colamosea (1975) demonstrated that leaders are in the best possible position to get things done if they first emerged informally as leaders in the group and then were elected by the members. Ben-Yoav, Hollander, and Carnevale (1983) similarly observed that elected leaders were more likely than are appointed leaders to contribute to the group’s discussion and to receive greater responsiveness and support subsequently from other members. Group members regard elected leaders as more competent, more responsive to the members’ needs, and more interested in the group’s task. Election gives followers a greater sense of responsibility for leadership and heightens their expectations for the leader’s performance (Hollander, 1978).
Legitimacy will allow leaders more latitude when representing their group in negotiations with higher authority or with other groups. In turn, legitimate leaders will expect to receive more support from the group they are representing (Boyd, 1972). If elected leaders are also self-confident, they are likely to use their legitimate power and persuasion instead of punishment to correct their subordinates’ poor performance (Goodstadt & Kipnis, 1970). Hollander, Fallon, and Edwards (1977) found that elected leaders had greater influence than appointed leaders on small experimental groups that were perceived by their members to be failing, although the greater influence of a newly elected leader in a group that continued to fail was soon lost.
Winning the election establishes a much higher degree of legitimate acceptance of the U.S. president as the embodiment of the nation and as the head of state, head of government, head of political party, and commander in chief of the military than would be expected from the president’s initial support from the voters. Usually, only about half of the eligible U.S. electorate votes in presidential elections. Often, the difference between the votes for the winner and the votes for the loser is so close that the president, with his enormous powers, is legitimated by only about 25% of the eligible voters. In 2000, the Republican candidate, George W. Bush, received 500,000 fewer popular votes than his Democratic opponent, Al Gore; but the Republican-appointed majority of justices of the U.S. Supreme Court decided 5–4 to award the contested 25 electoral votes of Florida to Bush to give him a majority of the electoral vote and thus legitimacy as president. Although for a while many Democrats continued to question Bush’s legitimacy, it was helped by Gore’s quickly announced concession of defeat.
Other experiments have shown that legitimacy is enhanced when an office is permanent rather than temporary and when it is filled by appointment by a higher authority rather than by election. As mentioned earlier, leaders’ negotiations with higher authorities will have more legitimacy in the eyes of group members if the leaders are elected by the groups rather than fortuitously appointed (Julian, Hollander, & Regula, 1969), but arbitrarily appointed leaders can increase their legitimacy by consulting with the members (Lamm, 1973).
Appointed leaders in a hierarchical setting also are likely to have legitimate power. The amount of power they have is a direct reflection of the power and status of the legitimatizing authority. For example, the commissions of U.S. Army officers confer powerful legitimacy; the appointment of a university chancellor by a board of trustees does likewise. However, when the legitimator is seen as incompetent and lacking in authority, the appointee’s power and ability will be questioned (Knight & Weiss, 1980). In Torrance’s (1954) study of decision making in established and newly formed groups, the members’ influence on decision making depended on their position in the power structure of the group and was stronger in permanent than in transitory groups. Furthermore, although Raven and French’s (1958a) examination of groups with elected and emergent leaders concluded that “the very occupation of a key position in a structure lends legitimacy to the occupant,” the authors also noted that elected leaders were better liked than emergent leaders and their influence was better accepted. Anderson, Karuza, and Blanchard (1977) observed that individuals who were elected to an undesirable leadership position had greater social power than those elected to a desirable position, but that appointment to either position made no difference. Julian, Hollander, and Regula (1969) found that the appointed leaders in experimental groups had a source of authority that was perceived to give them stability in their position. But election made the leaders more vulnerable to censure if they later proved inadequate. Members were most willing for the leaders to continue in their positions when the leaders were competent at the task, interested in the group members, and involved in the group’s activities.
Inner Circle versus Outer Circle. Vecchio (1979) showed that the leader’s legitimate power was particularly important to the performance of those subordinates who were not part of the leader’s “inner circle” of confidants and trusted assistants. At the same time, the subordinates in the inner circle did best when their superior did not possess power that was due to his or her formal position. The inner circle was most optimistic about its leader when the outer circle treated their leader in a friendly manner (Koulack, 1977).
Enhancement of Legitimate Power. Leaders can increase their legitimate power by becoming the accepted “makers of meaning,” definers of reality and the situation (Conger, 1991). In a formal organization, legitimate power becomes equivalent to the authority vested in a position. Illustrating the importance of such legitimate power, Klimoski, Friedman, and Weldon (1980) set up a laboratory simulation of an assessment center6 that required the assessors in a group meeting to integrate their assessments. The chairpersons of these integrative panel discussions either were or were not granted formal voting rights. If they were granted such rights, they were able to exert more influence on the panel and in the final evaluations of the candidates that emerged.
Subordinates’ trust of legitimate authority makes a difference in the authority’s power. Earley (1986b) reported that British tire workers who were assigned goals accepted them more when the reasoning for the goals was explained by their union steward, in whom they presumably felt more trust, than when the explanation was offered by their supervisors, whom they viewed as members of management—the untrustworthy opposition. The loss of such trust and confidence is particularly damaging to legitimate power.
Reduction in Legitimate Power. In Chapter 10, we noted the sharp fall in public trust in our business, government, and other institutions between 1945 and 2005 (Kingsbury, 2006). Mitchell and Scott (1987) examined the dramatic decline in the American public’s confidence in the legitimate leaders of various institutions. In the late 1970s, President Jimmy Carter saw it as a “crisis of confidence.” According to public opinion polls, trust in government declined from about 80% in the 1950s to 33% in 1976 (just after the Watergate scandal). From the late 1960s to the late 1970s, the American public’s confidence in business leaders fell equally dramatically. It was the leaders of government and business who were seen as particularly untrustworthy, not necessarily the institutions. In the 1980s, over 100 political appointees of the Reagan administration were forced to resign in a cloud of scandals and indictments. The “sleaze factor” in the Reagan administration did not help to restore confidence in governmental leaders; nor did the Wall Street scandals, the Iran-Contra affair, the many plant closings, or the stock market crash of 1987. Confidence in Ronald Reagan appeared to have been maintained during his administration because he conveyed the impression of a personable, likeable, strong, and decisive leader.
New and different scandals emerged during the presidencies of Bill Clinton and George W. Bush. Business leadership again suffered a crisis of confidence in the 1990s and the early twenty-first century from scandals, bankruptcies, the collapse of the boom in technology stocks, the huge expansion of the federal deficit, the continuing overseas migration of jobs, the decline in real income of the working class, the increasing gap between the rich and the poor, the lack of one-seventh of the population of health insurance, and the unpopular war in Iraq. Lipset (1985) found that a “great deal of confidence in the people running the various institutions” was expressed in 1966, but by 1984 it had declined, as follows: physicians, from 72% to 43%; educators, 61% to 40%; military leaders, 62% to 45%; religious leaders, 41% to 24%; members of the Supreme Court, 50% to 35%; CEOs of major companies, 55% to 19%; leaders of labor unions, 22% to 12%; members of Congress, 42% to 28%; and journalists, 29% to 18%. These figures continued to fall into the twenty-first century (Kingsbury, 2006).
The legitimacy of these institutional leaders is founded on the public’s belief in their special expertise, in their serving as stewards with legal and moral responsibility for the management of their institutions in the best interests of their constituencies, and in their ability to innovate and to inspire progress. Mitchell and Scott (1987) suggested that the decline in confidence was due to the loss of belief that institutional leaders had the expertise and motives that were once attributed to them. Rather than serving as stewards, the leaders had become self-serving. With the decline in confidence has come the loss of legitimacy. The competence of the leaders in government is increasingly questioned, particularly because of continued mishandling of terrorism and natural disasters. Progressive taxation has been bent out of shape, with the heaviest relative burden falling on the poor and working classes and increasingly favored treatment of the wealthiest. Leaders appear to be more self-aggrandizing. Their prospective innovations are increasingly constrained by the felt need to protect others from them. The decline in their legitimacy results in a public that feels much less “obligation or responsibility to believe in what their leaders tell them or to do what their leaders ask of them” (Mitchell & Scott, 1987, p. 449).
Countervailing power can also reduce legitimacy. Caplow (1968) observed that a majority coalition formed in opposition to leaders may not only undermine the legitimacy of their position but may take over the leadership. For instance, Worthy, Wright, and Shaw (1964) created groups in which a confederate accused a naïve teammate of losing a game. When the accusation was legitimate, other members of the team were less willing to interact with the accused member. When the accusation was not legitimate, they were less willing to interact with either the accused or the accuser.
Those with legitimate power may also be ready to give it up or to avoid exercising it. This tendency is more common in young low-level managers. In a survey of 569 managers, Veiga (1986) identified various causes for giving up control:
1. The manager, especially the young low-level manager, may think the group lacks direction, commitment, or a belief that it can do well.
2. The manager personally feels no commitment to or responsibility for the outcome and has nothing to gain from it.
3. The manager feels unwilling or unable to make a meaningful contribution to the decision.
4. The group’s task may seem ill-defined or too difficult.
5. The manager may believe that someone else in the group has more conviction, experience, ability, and willingness to take responsibility.
6. Someone else’s idea may be seen as similar to the manager’s.
7. The manager may feel intimidated by an attack or by higher authority.
8. The manager may wish to avoid damaging a relationship.
When Legitimacy Is Counterproductive. Kaplan, Drath, and Kofodimos (1985) noted that legitimate power may be counterproductive if the superior’s demeanor, isolation, and autonomy impede the flow of criticism from the subordinates to the superior. The subordinates’ performance may suffer if the leader’s legitimacy results in an aura of power, in a tendency to monopolize discussions, or in an abrasive style. The use of power in this way may increase social distance between the superior and the subordinates, further restricting communication. If subordinates perceive that the superior’s power is an embodiment of the organization, they may interpret criticism of the superior as disloyalty to the organization.
Legitimacy will be counterproductive if powerful superiors appoint only subordinates who agree with them or if the subordinates exempt themselves from the formal appraisals that others who are less powerful must undergo. But knowingly or unknowingly, supervisors may give up some of their legitimate power to their subordinates as a way of increasing overall satisfaction and responsibility. Blake and Mouton (1961a) asked a supervisor and a subordinate to rank lists of items in order of importance. The level of satisfaction and responsibility of both supervisors and subordinates combined was higher when they alternated in exercising power.7
Yukl and Falbe (1991) added to or replaced the five bases of power with charisma, persuasiveness, and information. Information is a particularly important base of power for political leaders. It gives political, congressional, and parliamentary leaders important knowledge about the rules of their legislatures, the predilections of key members, and their constituents’ needs and attitudes. Yasir Arafat’s continued presidency of the Palestinian Authority was based partly on his encyclopedic knowledge of the tribes and factions from the many towns, cities, and refugee camps whose diverse interests he had to reconcile if he was to maintain internal harmony (Dallas Morning News, 1997). For business leaders, accurate information about the informal networks in an organization of friends, potential advisors, and central figures can also be an important base of power (Krackhart, 1990).
Relative Importance. In 13 studies reviewed by Podsakoff and Schriesheim (1985a), large samples of respondents ranked French and Raven’s five bases of power according to their importance in the superior-subordinate relationship. Agreement was high among the respondents, regardless of whether they were salespeople, members of a liberal arts faculty, insurance agents, or factory workers. For each sample as a whole, expert and legitimate power always ranked first or second in importance. Mean ranks were tied at 1.6. The other bases were seen as far less important. Means were: referent, 3.3; reward, 4.1; coercive, 4.6. However, Podsakoff and Schriesheim urged caution in interpreting these findings on their face value.
Relative Willingness to Use Power. Effective principals establish higher standards for their school by using their power to enlist the support of the teachers to improve standards. They use their power to support the teachers’ career development during early adulthood to midlife and to help teachers find fulfillment in their work (Krupp, 1986). Leaders differ in their willingness to use power that derives from one base rather than another. Lord (1977) identified 12 functions that are typically performed by a leader, such as developing plans, proposing solutions, and providing resources. He watched to see how often each leader in a sample displayed the 12 functions. He also determined which of French and Raven’s (1959) five bases of power the leaders used in attempts to exert their power. Lord then correlated the occurrence of each leadership function with the basis of power the leader used. He found that task-relevant behavior to complete the work of the group correlated with the type of power used, but the leaders’ efforts to establish socioemotional relations did not. For example, the extent to which leaders relied on themselves as task experts correlated with the extent to which they proposed solutions. However, the extent to which the leaders tried to use referent power in their socioemotional efforts did not increase their being liked.
Coercion is more likely to be used by leaders who lack self-confidence and who have coercive power. In a laboratory experiment, Instone, Major, and Bunker (1983) found that, compared with those whose self-confidence was high, those who lacked self-confidence tended to use coercion rather than expert power. House (1984) listed other personal attributes that contribute to a leader’s tendency to use coercive power, including the leader’s power orientation, Machiavellianism, and dogmatism. In another laboratory study, Goodstadt and Hjelle (1973) reported that individuals who perceived themselves to be externally controlled by their environment used punishment to maintain their influence much more frequently than those who saw themselves as internally self-controlling. “Self-controlling internals” attempted to lead through their expert power rather than coercively.
Raven and Kruglanski (1970) suggested that those with power anticipate the consequences of using the various kinds of power they may have and avoid using powers that are believed to be least effective. Thus coercive power will be avoided unless the power holders expect resistance—in which case they will be tempted to employ more coercive power if they are capable of doing so (Kipnis, 1976). Support for this contention was found for managers in state agencies, industry, and the military (Kipnis & Cosentino, 1969), as well as in laboratory experiments (Goodstadt & Kipnis, 1970). In the Kipnis studies, supervisors were asked to describe an incident in which they had to correct the behavior of a subordinate. Which type of power was brought to bear in this situation depended on the supervisors’ diagnosis of the subordinate’s problem, the subsequent and most likely reactions of the subordinate, and what was likely to be best for the continuing relationship with the subordinate.
Culture, tradition, hierarchical relations, and organizational norms will also determine what type of power may be applied in a given situation. For instance, tradition may give those with seniority and age power that can be used and accepted by those who are younger and have less seniority. Age itself may make a difference. When I asked Japanese managers meeting with each other for the first time at a workshop to form informal groups with group leaders, the groups selected their leaders by asking members who was oldest. More authoritarian, coercive uses of power will come into greater use and acceptability during times of stress. Managers can use more reward and legitimate power with subordinates than with peers (Yukl, Kim, & Falbe, 1996).
Legitimacy will be stressed by those who believe in “rules of law rather than rules of men.” House (1984) suggested that legitimate power will be favored over arbitrariness and political influence because legitimate power provides more orderliness and predictability. Legitimacy has more clear-cut limitations; unbounded arbitrary power generates excess costs and inefficiencies. Legitimate power can be exerted impersonally without a buildup of allegations of political exchanges. Legitimate power is ordinarily more acceptable to the subordinate and more supported by the larger group than power based on the personal arbitrary predilections of the superior. House (1984) obtained considerable support for the greater acceptability of legitimacy in studies in which employees were asked why they complied with manager’s requests, suggestions, or directions.
Willingness to use one type of power rather than another was observed by Rosenberg and Pearlin (1962) in a study of the attitudes of 1,138 hospital nurses toward power. The percentages of nurses who stated that they would use each form of power were as follows: persuasion, 54%; benevolent manipulation, 38%; legitimate authority, 5%; coercive power, 2%; contractual power, 1%. Kappelman (1981) reported that male school principals were more likely than their female counterparts to use reward and coercive power, although women principals were likely to be described as more active leaders.
Since the bases of power are interrelated, changes in one will have effects on the others. For example, Greene and Podsakoff (1981) found that when the reward power of 37 supervisors in a paper mill was lowered by abandoning an incentive pay plan in which the supervisors’ evaluation of the performance of 392 subordinates had figured strongly, not only did the supervisors’ reward power decline, but their referent, legitimate, and organizationally sanctioned power declined as well, while the subordinates’ perceptions of the supervisors’ coerciveness increased significantly. No such changes in the bases of supervisory power were found in a second comparable paper mill that did not change its pay policy. Consistent with this, Garrison (1968) concluded that when leaders’ efforts to succeed through legitimacy failed, they resorted to coercive power.
The compliance caused by different bases of power is perceived differently. Rodrigues and Lloyd (1998) concluded from five studies using a total of 570 participants that compliance induced by reward, informational, and referent power was more internalized and controllable than compliance induced by expert, legitimate, and coercive power. Rubin, Lewicki, and Dunn (1973) found that cooperativeness, benevolence, friendliness, and generosity were attributed to those in power who used rewards rather than penalties. The reverse attributions were made about those who used authority or political persuasion.
Table 11.2 shows correlations between the perceived power base of first-line supervisors and their initiating behavior and consideration as leaders.8 Martin and Hunt (1980) found these results for 289 professional personnel in a construction bureau and 118 in a design bureau of a midwestern state highway department who responded to a mailed survey. As can be seen, of all the bases of power, expert power made the most important contribution to leadership. The effect was not always beneficial; design professionals who complied with their supervisor because of the latter’s expert power signaled a greater intention to quit.
Ivancevich (1970) reported that the satisfaction of life insurance agents correlated .35 with the agency manager’s use of referent power, expert power, and reward power, but satisfaction correlated 2.28 with the use of coercive power and .14 with the use of legitimate power. Bachman (1968) found that faculty members at colleges were better satisfied under powerful deans with control over college affairs whose influence was based on expert and referent power than under deans who relied on reward, legitimate, or coercive power. However, in a study of salesmen’s perceptions of branch managers’ power, Ivancevich and Donnelly (1970a) found that coercive and legitimate powers were not related to productivity. Only the expert and referent powers of the branch managers were positively related to the salesmen’s performance.
Caveat. Podsakoff and Schriesheim’s (1985a) review of these and other field studies of the relationship of the five sources of power of supervisors and the outcomes for subordinates concluded that unlike studies of the actual reward behavior of supervisors that usually generates positive outcomes, the supervisors’ power to reward was unrelated or even negatively related to such outcomes. But Podsakoff and Schriesheim attributed the difference in findings to the way these field studies measured power. Often the measure consisted only of respondents’ rank ordering of the bases of power as defined. The power to reward was confounded with a sense of illegitimacy used in exchange for reluctant compliance rather than for performance. Coerciveness was seen to be confounded with the potentially illegitimate use of punishment. The actual narrow content of the scales of legitimate power failed to match French and Raven’s broader conceptualization, and the desired identification of the subordinate with the superior was not included in measures of referent power, which often seemed no more than assessments of friendship.
Table 11.2 Correlations between the Perceived Basis of Power and Ratings of the Initiation of Structure and Consideration by First-Line Supervisors
Basis of Power |
Initiation of Structure* |
Consideration* |
||
Construction Bureau |
Design Bureau |
Construction Bureau |
Design Bureau |
|
Expert |
.44 |
.41 |
.50 |
.48 |
Referent |
.23 |
.16 |
.15 |
.10 |
Reward |
.24 |
.06 |
.11 |
.02 |
Coercive |
.10 |
.01 |
-.10 |
.01 |
Legitimate |
.04 |
.11 |
-.03 |
.00 |
* See Chapter 20 for detailed definitions of initiation and consideration.
SOURCE: Adapted from Martin and Hunt (1980).
Hinkin and Schriesheim’s (1989) surveys of three samples of students who described their part-time or full-time supervisors found that although coercive power correlated with legitimate power to some extent, it was likely to be independent of the other bases of power. The other bases—expert, referent, and reward—correlated with legitimate power as a base as well as with each other as bases.
Despite these shortcomings, seven of the field studies9 in Podsakoff and Schriesheim’s (1985a) review emerged consistently with positive associations between supervisors’ expert and referent power and subordinates’ satisfaction and performance. The results were uniformly negative or reversed for coercive power and mixed for legitimate and reward power. Negative correlations were found between the supervisors’ expert and referent power and the excused absences of production workers (Student, 1968), but no power measure was related to unexcused absences or turnover. However, Busch (1980) found that the expert power of supervisors contributed to the subordinates’ intentions to remain in the three companies studied. Less consistent findings emerged for referent and legitimate power. A similar pattern was noted by Gemmill and Thamhain (1974) in the outcomes of support by supervisors and the commitment of subordinates to the work associated with supervisory powers. Only expert power was consistently positively correlated with such outcomes and with the clarity of the subordinates’ role. Also, such role clarity appeared to be enhanced if the superiors had referent power.
Yukl and Taber (1983) agreed with Podsakoff and Schriesheim (1985a) that the use of expert and referent power is most efficacious. Nevertheless, the skill of application and appropriateness of power must be considered. Leaders need to exercise authority with courtesy and clarity and to verify compliance. Reward power should be used to reinforce desirable behaviors after they occur. Coercive power should be used only when absolutely necessary, for example, to deter behavior that is detrimental to the individual, group, or organization. Leaders accumulate and foster expert and referent power over time by showing that they are confident, decisive, considerate, and protective of subordinates’ interests. In the same vein, Watts (1986, p. 286) suggested that “using too much power can express the need to control, the flip side of which is the fear of being out of control. Using too little power can express a tremendous need to be liked, to avoid conflict.” Furthermore, leaders need to avoid promising more (on the basis of their reward power) than they can deliver, and they need to avoid being corrupted by their power.
In the 1840s, Count Cavour, a future cofounding statesman of modern Italy, wrote that “absolute power inevitably corrupts,” a remark which Lord Acton later made famous as “Power tends to corrupt; absolute power corrupts absolutely” (D. M. Smith, 1985, p. 26). The process was elucidated by Kipnis (1976). First, the desire for power becomes an end in itself. Next, access to power tempts the officeholder to use institutional resources for illegitimate self-aggrandizement. Self-aggrandizement is followed by false feedback from others, which elevates the leader’s self-esteem, devalues the worth of others, and distances the leader from them.
Gardner (1986a) listed four ways to help mitigate the corruption of power in government. First, the rule of law and the accountability of leaders to abide within explicit and universally applicable constraints must be accepted. Second, power needs to be dispersed, as in the U.S. system of checks and balances of executive, legislative, and judicial powers. Third, a strong private sector must be maintained. Fourth, an alert citizenry is essential.
So far, analyses of power and corruption have been left to investigative journalists and the legal profession. Controlled surveys and experiments are needed to increase our understanding of the interrelation of power-corruption relationship, its antecedents and its consequences for business, governmental, and other institutions. How do various factors—the fear of being caught; public humiliation; regulatory control; a sense of duty, obligation, patriotism, and loyalty; and internalized ethical and professional standards—inhibit the corruptive influence of holding power over others? Chapter 9, on the ethics of leadership, has already dealt with some aspects of the corrupt use of power.
Organizational formalization, rules, and norms can substitute for the expert power of leaders and their influence. On the other hand, if the expert power of leaders is seen to improve the functioning of their subordinates or to be applied usefully, their perceived influence will be enhanced (Podsakoff & Schriesheim, 1985). A leader’s referent power was found by Yukl and Falbe (1991) to be most important to compliance by followers; a leader’s legitimate and coercive power were most important for employee compliance in routine operations. If the holder of power is working on a task in an experiment with another person requiring cooperation, power will be used for their joint benefit. If they are in competition, power will be used to benefit its holder at the expense of the other. The amount of power used will not differ. The perceived competence of the power holder will affect the use of power in the cooperative context, but not in the competitive context (van Kippenberg, van Kippenberg, & Wilke, 2001).
Power is a force giving one person the potential to influence others. Power can be personal (expert and referent) and positional (reward, coercive, and legitimate). It contributes to leadership and influence but is not synonymous with them. Information is among many other sources of power. It makes power expandable. Status and esteem provide power, influence, and leadership. Followers react differently to leaders whose power derives from different sources. The followers seek to be liked by those with referent and reward power. Although the threat of punishment tends to induce compliance, followers find leaders who use coercive power less attractive than those who use other forms of power. Furthermore, followers respond to reward power either by developing contractual agreements or by forming coalitions that tend to equalize the bargaining positions of participants in the power structure. Appointment or election to a position tends to legitimate it to a greater degree than does acquisition of a position by force or by the emergence of a leader in interactions.
Leaders draw consciously or unconsciously on multiple sources of power. Reward and coercive power are controlled most easily. Leaders have less control over referent power—the extent to which group members like them and are attracted to them. Self-confidence, self-esteem, and knowledge of the task enable subordinates to resist the effects of power. Interacting with peers and gaining support from reference groups strengthen the followers’ resistance to coercion. The influence of a leader can be weakened if the followers form coalitions with counter-vailing power. Followers can reduce their subjection to influence by asking and receiving a clear definition of the situation. The power of a leader is weakened by the presence of members whose values and goals are in opposition to those of the leader and the organization and who challenge the legitimacy of the leadership role. Ingratiation may also be a means of private resistance. Power can corrupt both leaders and their followers. It can be shared and distributed for personal or organizational reasons.