Agency Theory

Academy ol Managwneni Review. 1989. VoL 14. No. i, S7-7i

Agency Theory: An Assessmentand Review

KATHLEEN M. EISENHARDTStanford University

Agency theory is an important, yet controversial, theory. This paperreviews agency theory, its contributions to organization theory, andthe extant empirical work and develops testable propositions. Theconclusions are that agency theory (a) offers unique insight into in-formation systems, outcome uncertainty, incentives, and risk and (b)is an empirically valid perspective, particularly when coupled withcomplementary perspectives. The principal recommendation is to in-corporate an agency perspective in studies of the many problemshaving a cooperative structure.

One day Deng Xiaoping decided to take hisgrandson to visit Mao. "Call me granduncle,"Mao offered warmly. "Oh, 1 certainly couldn'tdo that. Chairman Mao," the awe-struck childreplied. "Why don't you give him an apple?"suggested Deng. No sooner had Mao done sothan the boy happily chirped, "Oh thank you.Granduncle." "You see," said Deng, "what in-centives can achieve." ("Capitalism," 1984, p.62)

Agency theory has been used by scholars inaccounting (e.g., Demski & Feitham, 1978), eco-nomics (e.g., Spence & Zeckhauser, 1971), fi-nance (e.g., Fama, 1980), marketing (e.g., Basu,Lai, Srinivasan, & Staelin, 1985), political sci-ence (e.g., Mitnick, 1986), organizational behav-ior (e.g., Eisenhardt, 1985, 1988; Kosnik, 1987),and sociology {e.g., Eccles, 1985; White, 1985).Yet, it is still surrounded by controversy. Its pro-ponents argue that a revolution is at hand andthat "the foundation for a powerful theory of or-ganizations is being put into place" (Jensen,1983, p. 324). Its detractors call it trivial, dehu-manizing, and even "dangerous" (Perrow, 1986,p. 235).

Which is it: grand theory or great sham? The

purposes of this paper are to describe agencytheory and to indicate ways in which organiza-tional researchers can use its insights. The pa-per is organized around four questions that aregermane to organizational research. The firstasks the deceptively simple question. What isagency theory? Often, the technical style, math-ematics, and tautological reasoning of th©agency literature can obscure the theory. More-over, the agency literature is split into twocamps (Jensen, 1983), leading to differences ininterpretation. For example, Barney and Ouchi(1986) argued that agency theory emphasizeshow capital markets can affect the firm,whereas other authors made no reference tocapital markets at all (Anderson, 1985; Demski &Feitham, 1978; Eccles, 1985; Eisenhardt, 1985).

The second question is. What does agencytheory contribute to organizational theory? Pro-ponents such as Ross (1973, p. 134) argued that"examples of agency are universal." Yet otherscholars such as Perrow (1986) claimed thatagency theory addresses no clear problems,and Hirsch and Friedman (1986) called it exces-sively narrow, focusing only on stock price. Foreconomists, long accustomed to treating the or-

57

ganizatjon as a "black box" in the theory of thefirm, agency theory may be revolutionary. Yet,for organizational scholars the worth of agencytheory is not so obvious.

The third question is. Is agency theory empir-ically valid? The power of the empirical researchon agency theory to explain organizational phe-nomena is important to assess, particularly inlight of the criticism that agency theory is"hardly subject to empirical test since it rarelytries to explain actual events" (Perrow, 1986, p.224). Perrow (1986) also criticized the theory forbeing un realistically one-sided because of itsneglect of potential exploitation of workers.

The final question is, What topics and contextsare fruitful for organizational researchers whouse agency theory? Identifying how usefulagency theory can be to organizational scholarsrequires understanding the situations in whichthe agency perspective can provide theoreticalleverage.

The principal contributions of the paper are topresent testable propositions, identify contribu-tions of the theory to organizational thinking,and evaluate the extant empirical literature. Theoverall conclusion is that agency theory is a use-ful addition to organizational theory. Theagency theory ideas on risk, outcome uncer-tainty, incentives, and information systems arenovel contributions to organizational thinking,and the empirical evidence is supportive of thetheory, particularly when coupled with comple-mentary theoretical perspectives.

Origins of Agency Theory

During the 1960s and early 1970s, economistsexplored risk sharing among individuals orgroups (e,g.. Arrow, 1971; Wilson, 1968). Thisliterature described the risk-sharing problem asone that arises when cooperating parties havedifferent attitudes toward risk. Agency theorybroadened this risk-sharing literature to includethe so-called agency problem that occurs whencooperating parties have different goals and di-

vision of labor (Jensen & Meckling, 1976; Ross,1973). Specifically, agency theory is directed atthe ubiquitous agency relationship, in whichone party (the principal) delegates work to an-other (the agent), who performs that work.Agency theory attempts to describe this relation-ship using the metaphor of a contract (Jensen &Meckling, 1976).

Agency theory is concerned with resolvingtwo problems that can occur in agency relation-ships. The first is the agency probl&m that ariseswhen (a) the desires or goals of the principal andagent conflict and (b) it is difficult or expensivefor the principal to verify what the agent is ac-tually doing. The problem here is that the prin-cipal cannot verify that the agent has behavedappropriately. The second is the problem of risksharing that arises when the principal andagent have different attitudes toward risk. Theproblem here is that the principal and the agentmay prefer different actions because of the dif-ferent risk preferences.

Because the unit of analysis is the contractgoverning the relationship between the princi-pal and the agent, the focus of the theory is ondetermining the most efficient contract govern-ing the principal-agent relationship given as-sumptions obout people (e.g., self-interest,bounded rationality, risk aversion), organiza-tions (e.g., goal conflict among members), andinformation (e.g., information is a commoditywhich can be purchased). Specifically, thequestion becomes. Is a behavior-oriented con-tract (e.g., salaries, hierarchical governance)more efficient than an outcome-oriented con-tract (e.g., commissions, stock options, transferof property rights, market governance)? An over-view of agency theory is given in Table 1.

The agency structure is applicable in a varietyof settings, ranging from macrolevel issues suchas regulatory policy to microlevel dyad phe-nomena such as blame, impression manage-ment, lying, and other expressions of self-interest. Most frequently, agency theory hasbeen applied to organizational phenomena

58

Table 1Agency Theory Overview

Key idea

Unit ofanalysis

Humanassumptions

Organizationalassumptions

informationassumption

Contractingproblems

Problemdomain

Principal-agent relationships shouldreflect efficient organization ofinformation and risk-bearing costs

Contract between principal and agent

Seif-interestBounded rationality-Risk aversion

Partial goal conflict among participantsEfficiency as the effectiveness criterionInformation asymmetry between principal

and agent

Infonnation as a purchasable commodity

Agency (moral hazard and adverseselection)

Risk shanng

Relationships in which the principal andagent have partly differing goals andrisk preferences (e.g., compensation,regulation, leadership, impressionmanagement, whistle-blowing, verticalintegration, transfer pricing)

such as compensation (e.g., Conlon & Parks,1988; Eisenhardt, 1985), acquisition and diversi-fication strategies (e.g., Amihud & Lev, 1981),board relationships (e.g., Fama & Jensen, 1983;Kosnik, 1987), ownership and financing struc-tures (e.g., Argawal & Mandelker, 1987; Jensen& Meckling, 1976), vertical integration (Ander-son, 1985; Eccles, 1985), and innovation (Bolton,1988; Zenger, 1988). Overall, the domain ofagency theory is relationships that mirror thebasic agency structure of a principal and anagent who are engaged in cooperative behav-ior, but have differing goals and differing atti-tudes toward risk.

Agency TheoryFrom its roots in information economics,

agency theory has developed along two lines:

positivist and principal-agent Censen, 1983). Thetwo streams share a common unit of analysis:the contract between the principal and theagent. They also share common assumptionsabout people, organizations, and information.However, they differ in their mathematical rigor,dependent variable, and style.

Positivist Agency Theory

Positivist researchers have focused on identi-fying situations in which the principal and agentare likely to have conflicting goals and then de-scribing the governance mechanisms that limitthe agent's self-serving behavior. Positivist re-search is less mathematical than principal-agent research. Also, positivist researchershave focused almost exclusively on the specialcase of the principal-agent relationship betweenowners and managers of large, public corpora-tions (Berle & Means, 1932).

Three articles have been particularly influen-tial. Jensen and Meckling (1976) explored theownership structure of the corporation, includ-ing how equity ownership by managers alignsmanagers' interests with those of owners. Fama(1980) discussed the role of efficient capital andlabor markets as information mechanisms thatare used to control the self-serving behavior oftop executives. Fama and Jensen (1983) de-scribed the role of the board of directors as aninformation system that the stockholders withinlarge corporations could use to monitor the op-portunism of top executives. Jensen and his col-leagues (Jensen, 1984; Jensen & Roeback, 1983)extended these ideas to controversial practices,such as golden parachutes and corporate raid-ing.

From a theoretical perspective, the positiviststream has been most concerned with describ-ing the governance mechanisms that solve theagency problem. Jensen (1983, p. 326) describedthis interest as "why certain contractual rela-tions arise." Two propositions capture the gov-ernance mechanisms which are identified in thepositivist stream. One proposition is that out-

59

come-based contracts are effective in curbingagent opportunism. The argument is that suchcontracts coalign the preferences of agents withthose of the principal because the rewards forboth depend on the same actions, and, there-fore, the conflicts of self-interest between princi-pal and agent are reduced. For example, Jensenand Meckiing (1976) described how increasingthe firm ownership of the managers decreasesmanagerial opportunism. In formal terms.

Proposition I: When the contract between theprincipal and agent is oufcome based, (heagent is more likely to behave in the interests oftbe principal.

The second proposition is that infonnation sys-tems also curb agent opportunisn:i. The argu-ment here is that, since information systems in-form the principal about what the agent is actu-ally doing, they are likely to curb agent oppor-tunism because the agent will realize that he orshe cannot deceive the principal. For example,Fama (1980) described the informcrtion effects ofefficient capital and labor markets on manage-rial opportunism, and Fama and Jensen (1983)described the information role that boarc^ of di-rectors play in controlling managerial behavior.In formal terms.

Proposition 2: When the principal has informa-tion to verify agent behavior, the agent is morelikely io behave in the interests of the principal

At its best, positivist agency theory can be re-garded as enriching economics by offering amore complex view of organizations (Jensen,1983). However, it has been criticized by orga-nizational theorists as minimalist (Hirsch,Michaels, & Friedman, 1987; Perrow, 1986) andby microeconomists as tautological and lackingrigor (Jensen, 1983). Nonetheless, positivistagency theory has ignited considerable re-search (Barney & Ouchi, 1986) and popular in-terest ("Meet Mike," 1988).

Principal-Agent ReBearch

Principal-agent researchers are concernedwith a general theory of the principal-agent re-lationship, a theory that can be applied to em-ployer-employee, lawyer-client, buyer-supplier,and ether agency relationships (Harris & Raviv,1978). Characteristic oi formal theory, the prin-cipal-agent paradigm involves careful specifi-cation of assumptions, which are followed bylogical deduction and mathematical proof.

In comparison with the positivist stream, prin-cipal-agent theory is abstract and mathematicaland, therefore, less accessible to organizationalscholars. Indeed, the most vocal critics of thetheory (Perrow, 1986; Hirsch et al., 1987) havefocused their attacks primarily on the morewidely known positivist stream. Also, the prind-pal-agent stream has a broader focus andgreater interest in general, theoretical implica-tions. In contrast, the positivist writers have fo-cused almost exclusively on the special case ofthe owner/CEO relationship in the large corpo-ration. Finally, principal-agent research in-cludes many more testable implications.

For organizational scholars, these differencesprovide background for understanding criticismof the theory. However, they are not crucialRather, the important point is that the twostreams are complementary: Positivist theoryidentifies various contract alternatives, and prin-cipal-agent theory indicates which contract isthe most efficient under varying levels of out-come uncertainty, risk aversion, information,and other variables described below.

The focus of the principal-agent literature ison determining the optimal contract, behaviorversus outcome, between the principal and theagent. The simple model assumes goal conflictbetween principal and agent, an easily mea-sured outcome, and an agent who is more riskaverse than the principal. (Note: The argumentbehind a more risk averse agent is that agents,who are unable to diversify their employment,should be risk averse and principals, who are

60

capable of diversifying their investments,should be risk neutral.) The approach of the sim-ple model can be described in terms of cases(e.g., Demski 8c Feitham, 1978). The first case, asimple case of complete information, is when theprincipal knows what the agent has done.Given that the principal is buying the agent'sbehavior, then a contract that is based on be-havior is meet efficient. An outcome-based con-tract would needlessly transfer risk to the agent,who is assumed to be more risk averse than theprincipal.

The second case is when the principal doesnot know exactly what the agent has done.Given the self-interest of the agent, the agentmay or may not have behaved as agreed. Theagency problem arises because (a) the principaland the agent have different goals and (b) theprincipal cannot determine if the agent has be-haved appropriately. In the formal literature,two aspects of the agency problem are cited.Moral hazard refers to lack of effort on the part ofthe agent. The argument here is that the agentmay simply not put forth the agreed-upon effort.That is, the agent is shirking. For example,moral hazard occurs when a research scientistworks on a personal research project on com-pany time, but the research is so complex thatcorporate management cannot detect what thescientist is actually doing. Adverse selection re-fers to the misrepresentation of ability by theagent. The argument here is that the agent mayclaim to have certain skills or abilities when heor she is hired. Adverse selection arises becausethe principal cannot completely verify theseskills or abilities either at the time of hiring orwhile the agent is working. For example, ad-verse selection occurs when a research scientistclaims to have experience in a scientific spe-cialty and the employer cannot judge whetherthis is the case.

In the case of unobservable behavior (due tomoral hazard or adverse selection), the principalhas two options. One is to discover the agent'sbehavior by investing in information systems

such as budgeting systems, reporting proce-dures, boards of directors, and additional layersof management. Such investments reveal theagent's behavior to the principal, and the situa-tion reverts to the complete information case. Informal terms,

Proposition 3: Information systems are posi-tively related to behavior-based contracts andnegatively related to oufcome-based contracts.

The other option is to contract on the outcomesof the agent's behavior. Such an outcome-basedcontract motivates behavior by coalignment ofthe agent's preferences with those of the princi-pal, but at the price of transferring risk to theagent. The issue of risk arises because outcomesare only partly a function of behaviors. Govem-ment policies, economic climate, competitor ac-tions, technological change, and so on, maycause uncontrollable variations in outcomes.The resulting outcome uncertainty introducesnot only the inability to preplan, but also riskthat must be borne by someone. When outcomeuncertainty is low, the costs of shifting risk to theagent are low and outcome-based contracts areattractive. However, as uncertainty increases, itbecomes increasingly expensive to shift risk de-spite the motivational benefits of outcome-basedcontracts. In formal terms.

Proposition 4: Outcome uncertainty is positivelyrelated to behavior-based contracts and nega-tively related to outcome-based contracts.

This simple agency model has been describedin varying ways by many authors (e.g., Demski& Feitham, 1978; Harris & Raviv, 1979; Holm-strom, 1979; Shavell, 1979). However, the heartof principal-agent theory is the trade-off be-tween (a) the cost of measuring behavior and (b)the cost of measuring outcomes and transferringrisk to the agent.

A number of extensions to this simple modelare possible. One is to relax the assumption of arisk-averse agent (e.g., Harris & Raviv, 1979).Research (MacCrimmon & Wehrung, 1986) indi-cates that individuals vary widely in their risk

61

attitudes. As the agent becomes increasinglyless risk averse (e.g., a wealthy agent), it be-comes more attractive to pass risk to the agentusing an outcome-based contract. Conversely,as the agent becomes more risk averse, it is in-creasingly expensive to pass risk to the agent. Informal terms.

Proposition 5; Th& risk aversion of the agent ispositively related to behavior-based contractsand negatively related to outcome-based con-tracts.

Similarly, as the principal becomes more riskaverse, it is increasingly attractive to pass risk tothe agent. In formal terms.

Proposition 6: The risk aversion ot the principalis negatively related to behavior-based con-tracts and positiveJy related to outcome-based contracts.

Another extension is to relax the assumptionof goal conflict between the principal and agent(e.g., Demski, 1980). This might occur either in ahighly socialized or clan-oriented firm (Ouchi,1979) or in situations in which self-interest givesway to selfless behavior (Perrow, 1986). If thereis no goal conflict, the agent will behave as theprincipal would like, regardless of whether hisor her behavior is monitored. As goal conflictdecreases, there is a decreasing motivationalimperative for outcome-based contracting, andthe issue reduces to risk-sharing considerations.Under the assumption of a risk-averse agent,behavior-based contracts become more attrac-tive. In formal terms,

Proposition 7: The goal conflict between princi-pal and agent is negatively related to behavior-based contracts and positively related to out-come-based contracts.

Another set of extensions relates to the task per-formed by the agent. For example, the progam-mability of the task is likely to influence the easeof measuring behavior (Eisenhardt, 1985, 1988).Programmability is defined as the degree towhich appropriate behavior by the agent canbe specified in advance. For example, the job of

a retail sales cashier is much more programmedthan that of a high-technology entrepreneur.The argument is that the behavior of agents en-gaged in more programmed jobs is easier to ob-serve and evaluate. Therefore, the more pro-grammed the task, the more attractive are be-havior-based contracts because informationabout the agent's behavior is more readily de-termined. Very programmed tasks readily re-veal agent behavior, and the situation reverts tothe complete information case. Thus, retail salesclerks are more likely to be paid via behavior-based contracting (e.g., hourly wages), where-as entrepreneurs are more likely to be compen-sated with outcome-based contracts (e.g., stockownership). In formal terms.

Proposition 8: TasJr programmabiiity is posi-..tiveiy reiated to behavior-based contracts andnegatively related to outcome-based contracts.

Another task characteristic is the measurabil-ity of the outcome (Anderson, 1985; Eisenhardt,1985). The simple model assumes that outcomesare easily measured. However, some tasks re-quire a long time to complete, involve joint orteam effort, or produce soft outcomes. In thesecircumstances, outcomes are either difficult tomeasure or difficult to measure within a practi-cal amount of time. When outcomes are mea-sured with difficulty, outcome-based contractsare less attractive. In contrast, when outcomesare readily measured, outcome-based contractsare more attractive. In formal terms.

Proposition 9: Outcome measurability is nega-tively related to behavior-based contracts andpositively related to outcome-based contracts.

Finally, it seems reasonable that when prin-cipals and agents engage in a long-term rela-tionship, it is likely that the principal will leamabout the agent (e.g., Lambert, 1983) and so willbe able to assess behavior more readily. Con-versely, in short-term agency relationships, theinformation asymmetry between principal andagent is likely to be greater, thus making out-

come-based contracts more attractive. In formaltenns,

Proposition 10: The length of the agency rela-tionship is positively related to behavior-basedcontracts and negatively related to outcome-based contracts.

Agency Theory and theOrganizational Literature

Despite Perrow's (1986) assertion that agencytheory is very different from organization theory,agency theory has several links to mainstreamorganization perspectives (see Table 2). At itsroots, agency theory is consistent with the clas-sic works of Barnard (1938) on the nature of co-operative behavior and March and Simon (1958)on the inducements and contributions of the em-ployment relationship. As in this earlier work,the heart of agency theory is the goal conflictinherent when individuals with differing prefer-ences engage in cooperative effort, and the es-sential metaphor is that of the contract.

Agency theory is also similar to political mod-els of organizations. Both agency and politicalperspectives assume the pursuit of self-interestat the individual level and goal conflict at theorganizational level (e.g., March, 1962; Pfeffer,1981). Also, in both perspectives, infonnation

asymmetry is linked to the power of lower orderparticipants (e.g., Pettigrew, 1973). The differ-ence is that in political models goal conflicts areresolved through bargaining, negotiation, andcoalitions—the power mechanism of politicalscience. In agency theory they are resolvedthrough the coalignment of incentives—theprice mechanism of economics.

Agency theory also is similar to the informa-tion processing approaches to contingency the-ory (Chandler, 1962; Galbraith, 1973; Lawrence& Lorsch, 1967). Both perspectives are informa-tion theories. They assume that individuals areboundedly rational and that information is dis-tributed asymmetrically throughout the organi-zation. They also are efficiency theories; that is,they use efficient processing of information as acriterion for choosing among various organizingforms (Galbraith, 1973). The difference betweenthe two is their focus: In contingency theory re-searchers are concerned with the optimal struc-turing of reporting relationships and decision-making responsibilities (e.g., Galbraith, 1973;Lawrence & Lorsch, 1967), whereas in agencytheory they are concerned with the optimalstructuring of control relationships resultingfrom these reporting and decision-making pat-terns. For example, using contingency theory,we would be concerned with whether a firm isorganized in a divisional or matrix structure.

Table 2Comparison of Agency Theory Assumptions and Organizational Perspectives

Assumption

Seli-interestGoal conflictBounded rationalityIniormation asymmetryPreemmence of efficiencyRisk aversionWormation as a commodity

PoliUcal

XX

Contingency

XXX

Perspective

OrganisationControl

X

X

TransactionCost

XXXXX

Agency

XXXXXXX

63

Using agency theory, we would be concernedwith whether managers within the chosen struc-ture are compensated by performance incen-tives.

The most obvious tie is with the organizationalcontrol literature (e.g., Dombusch Be Scott, 1974).For example, Thompson's (1967) and later Ou-chi's (1979) Unking of known means/ends rela-tionships and crystallized goals to behavior ver-sus outcome control is very similar to agencytheory's Unking task programmabiUty and mea-surability of outcomes to contract form (Eisen-hardt, 1985). That is, known means/ends rela-tionships (task programmability) lead to behav-ior control, and crystallized goals (measurableoutcomes) lead to outcome control. Similarly,Ouchi's (1979) extension of Thompsons (1967)framework to include clan control is similar toassuming low goal conflict (Proposition 7) inagency theory. Clan control impUes goal con-gruence between people and, therefore, the re-duced need to monitor behavior or outcomes.Motivation issues disappear. The major differ-ences between agency theory and the organi-zational control literature are the risk implica-tions of principal and agent risk aversion andoutcome uncertainty (Propositions A, b, 6).

Not surprisingly, agency theory has similari-ties with the transaction cost perspective(Williamson, 1975). As noted by Bomey and Ou-chi (1986), the theories share assumptions of self-interest and bounded rationality. They alsohave similar dependent variables; that is, hier-archies roughly correspond to behavior-basedcontracts, and markets correspond to outcome-based contracts. However, the two theoriesarise from different traditions in economics(Spence, 1975): In transaction cost theorizing weare concerned with organizational boundaries,whereas in agency theorizing the contract be-tween cooperating parties, regardless of bound-ary, is highlighted. However, the most impor-tant difference is that each theory includesunique independent variables. In transactioncost theory these are asset specificity and small

numbers bargaining. In agency theory thereare the risk attitudes of the principal and agent,outcome uncertainty, and information systems.Thus, the two theories share a parentage in eco-nomics, but each has its own focus and severalunique Independent variables.

Contributions of Agency Theory

Agency theory reestablishes the importanceof incentives and seif-interest in organizaUona]thinking (Perrow, 1986). Agency theory remindsus that much of organizational life, whether welike it or not, is based on self-interest. Agency-theory also emphasizes the importance of acommon problem structure across research top-ics. As Barney and Ouchi (1986) described it,organization research has become increasinglytopic, rather than theory, centered. Agency the-ory reminds us that common problem structuresdo exist across research domains. Therefore, re-sults from one research area (e.g., vertical inte-gration) may be germane to others with a com-mon problem structure (e.g., compensation).

Agency theory also makes two specific contri-butions to organizational thinking. The first is thetreatment of information. In agency theory, in-formation is regarded as a commodity: It has acost, and it can be purchased. This gives animportant role to formal information systems,such as budgeting, MBO, and boards of direc-tors, and informal ones, such as managerialsupervision, which is unique in organizationalresearch. The implication is that organizationscan invest in information systems in order tocontrol agent opportunism.

An illustration of this is executive compensa-tion. A number of authois in this literature haveexpressed surprise at the lack of performance-based executive compensation (e.g., Pearce,Stevenson, & Perry, 1985; Ungson & Steers,1984). However, from an agency perspective, ilis not surprising since such compensationshould be contingent upon a variety of factorsincluding information systems. Specifically,

64

richer infonnation systems control managerialopportunism and, therefore, lead to less perfor-mance-contingent pay.

One particularly relevant information systemfor monitoring executive behaviors is the boardof directors. From an agency perspective, boardscan be used as monitoring devices for share-holder interests (Fama & Jensen, 1983). Whenboards provide richer information, compensa-tion is less likely to be based on firm perfor-mance. Rather, because the behaviors of top ex-ecutives are better known, compensation basedon knowledge of executive behaviors is morelikely. Executives would then be rewarded fortaking well-conceived actions (e.g., highrisk/high potential R&D) whose outcomes maybe unsuccessful. Also, when boards providericher information, top executives are morelikely to engage in behaviors that are consistentwith stockholders' interests. For example, froman agency viewpoint, behaviors such as usinggreenmail and golden parachutes, which tendto benefit the manager more than the stockhold-ers, are less likely when boards are better mon-itors of stockholders' interests. Operationally,the richness of board information can be mea-sured in terms of characteristics such as fre-quency of board meetings, number of boardsubcommittees, number of board members withlong tenure, number of board members withmanagerial and industry experience, and num-ber of board members representing specificownership groups.

A second contribution of agency theory is itsrisk implications. Organizations are assumed tohave uncertain futures. The future may bringprosperity, bankruptcy, or some intermediateoutcome, and that future is only partly controlledby organization members. Environmental ef-fects such as government regulation, emer-gence of new competitors, and technical Inno-vation can affect outcomes. Agency theory ex-tends organizational thinking by pushing theramifications of outcome uncertainty to theirimplications for creating risk. Uncertainty is

viewed in terms of risk/reward trade-offs, not justin terms of inability to preplan. The implicationis that outcome uncertainty coupled with differ-ences in willingness to accept risk should influ-ence contracts between principal and agent.

Vertical integration provides an illustration.For example. Walker and Weber (1984) foundthat technological and demand uncertainty didnot affect the "make or buy" decision for compo-nents in a large automobile manufacturer (prin-cipal in this case). The authors were unable toexplain their results using a transaction costframework. However, their results are consistentwith agency thinking if the managers of the au-tomobile firm are risk neutral (a reasonable as-sumption given the size of the automobile firmrelative to the importance of any single compo-nent). According to agency theory, we wouldpredict that such a risk-neutral principal is rela-tively uninfluenced by outcome uncertainty,which was Walker and Weber's result.

Conversely, according to agency theory, thereverse prediction is true for a new venture. Inthis case, the firm is small and new, and it haslimited resources available to it for weatheringuncertainty: The likelihood of failure loomslarge. In this case, the managers of the venturemay be risk-averse principals. If so, according toagency theory we would predict that such man-agers will be very sensitive to outcome uncer-tainty. In pariicuiar, the managers would bemore likely to choose the "buy" option, therebytransferring risk to the supplying firm. Overall,agency theory predicts that risk-neutral manag-ers are likely to choose the "make" option (be-havior-based contract), whereas risk-averse ex-ecutives are likely to choose "buy" (outcome-based contract).

Empirical Results

Researchers in several disciplines have un-dertaken empirical studies of agency theory.These studies, mirroring the two streams of theo-retical agency research, are in Table 3.

65

66

3

CO

O

Isi<o o•^ bo 53 50 3

I - CO

II

lu

2 tai2 „, " ffi gS -S ° « c

IC/3

I

1(X,

a

f

67

Results of the Positivist Stream

In the positivisf stream, the common approachis to Identify a policy or behavior in which stock-holder and management interests diverge andthen to demonstrate that information systems oroutcome-based incentives solve the agencyproblem. That is, these mechanisms coaUgnmanagerial behaviors with owner preferences.Consistent with the positivist tradition, most ofthese studies concern the separation of owner-ship from management in large corporations,and they use secondary source data that arecxvailable ior large firms.

One of the earliest studies of this type wasconducted by Amihud and Lev (1981). These re-searchers explored why firms engage in con-glomerate mergers. In general, conglomeratemergers are not in the interests of the stockhold-ers because, typically, stockholders can diver-sify directly through their stock portfolio. In con-trast, conglomerate mergers may be attractiveto manageTS who have fewer avenues availableto diversify their own risk. Hence, conglomeratemergers are an arena in which owner and man-ager interests diverge. Specifically, these au-thors linked merger and diversification behav-iors to whether the firm was owner controlled (i.e.,had a major stockholder) or manager controlled(i.e., had no major stockholder). Consistent withagency theory arguments (Jensen & Meckling,1976), manager-controlled firms engaged in sig-nificantly more conglomerate (but not more re-lated) acquisitions and were more diversified.

Along the same lines. Walking and Long(1984) studied managers' resistance to takeoverbids. Their sample included 105 large U.S. cor-porations that were targets of takeover attemptsbetween 1972 and 1977. In general, resistance totakeover bids is not in the stockholders' interests,but it may be in the interests of managers be-cause they can lose their jobs during a takeover.Consistent with agency theory Qensen & Meck-Ung, 1976), the authors found that managerewho have substantial equity positions within

their firms (outcome-based contracts) were lesslikely to resist takeover bids.

The effects of market discipline on agency re-lationships were examined in Wolfson's (1985)study of the relationship between the limited(principals) and general (agent) partners in oiland gas tax shelter programs. In this study, bothtax and agency effects were combined in orderto assess why the limited partnership gover-nance form survived in this setting despite ex-tensive information advantages and divergentincentives for the limited partner. Consistentwith agency arguments (Fama, 1980), Wolfsonfound that long-run reputation effects of the mar-ket coallgned the short-run behaviors of the gen-eral partner with the limited partners' welfare.

Kosnik (1987) examined another informationmechanism for managerial opportunism, theboard of directors. Kosnik studied 110 large U.S.corporations that were greenmail targets be-tween 1979 and 1983. Using both hegemony andagency theories, she related board characteris-tics to whether greenmail was actually paid(paying greenmail is considered not in the stock-holders' interests). As predicted by agency the-ory (Fama & Jensen, 1983), boards of companiesthat resisted greenmail had a higher proportionof outside directors and a higher proportion ofoutside director executives.

In a similar vein, Argawal and Mandelker(1987) examined whether executive holdings offirm securities reduced agency problems be-tween stockholders and management. Specifi-cally, they studied the relationship betweenstock and stock option holdings of executivesand whether acquisition and financing deci-sions were made consistent with the interests ofstockholders. In general, managers prefer lowerrisk acquisitions and lower debt financing (seeArgawal & Mandelker, 1987, for a review). Theirsample included 209 firms that participated inacquisitions and divestitures between 1974 and1982. Consistent with agency ideas (e.g., Jensen& Meckling, 1976), executive security holdings(outcome-based contract) were related to acqm-

sition and financing decisiorxs that wereconsistent with stockholder interest. That is, ex-ecutive stock holdings appeared to coalignmanagerial preferences with those of stockhold-ers.

Singh and Harianto (in press) studied goldenparachutes in a matched sample of 84 Fortune500 firms. Their study included variables fromboth agency and managerialist perspectives.Consistent with agency theory Gensen & Meck-ling, 1976; Fama & Jensen, 1983), the authorsfound that golden parachutes are used to coaUgnexecutive interests with those of stockholders intakeover situations, and they are seen as an al-temative outcome-based contract to executivestock ownership. Specifically, the authors foundthat golden parachutes were positively associ-ated with a higher probability of a takeover at-tempt and negatively associated with executivestock holdings.

Finally, Barney (1988) explored whether em-ployee stock ownership reduces a firm's cost ofequity capital. Consistent with agency theoryGensen & Meckling, 1976), Barney argued thatemployee stock ownership (outcome-based con-tract) would coaUgn the interests of employeeswith stockholders. Using efficient capital marketassumptions, he further argued that this coaUgn-ment would be reflected in the market through alower cost of equity. Although Barney did notdirectly test the agency argument, the resultsare consistent with an agency view.

In summary, there is support for the existenceof agency problems between shareholders andtop executives across situations in which theirinterests diverge—that is, takeover attempts,debt versus equity financing, acquisitions, anddivestitures, and for the mitigation of agencyproblems (a) through outcome-based contractssuch as golden parachutes (Singh & Harianto,m press) and executive stock holdings (Argawal& Mandelker, 1987; Walking & Long, 1984) and(b) through information systems such as boards(Kosnik, 1987) and efficient markets (Barney,1988; Wolfson, 1985). OveraU, these studies sup-

port the positivist propositions described earlier.Similarly, laboratory studies by Dejong and col-leagues (1985), which are not reviewed here,are also supportive.

Results of the Principal-Agent Stream

The principal-agent stream is more directly fo-cused on the contract between the principal andthe agent. Whereas the positivist stream lays thefoundation (that is, that agency problems existand that various contract alternatives are avail-able), the principal-agent stream indicates themost efficient contract altemative in a given sit-uation. The common approach in these studiesis to use a subset of agency variables such astask programmability, information systems, andoutcome unceriainty to predict whether the con-tract is behavior- or outcome-based. The under-lying assumption is that principals and agentswill choose the most efficient contract, althoughefficiency is not directly tested.

In one study, Anderson (1985) probed verticalintegration using a transaction cost perspectivewith agency variables. Specifically, she exam-ined the choice between a manufacturer's rep-resentative (outcome-based) and a corporatesales force (behavior-based) among a sample ofelectronics firms. The most powerful explana-tory variable was from agency theory: the diffi-culty of measuring outcomes (measured byamount of nonselling tasks and joint team sales).Consistent with agency predictions, this vari-able was positively related to using a corporatesales force (behavior-based contract).

In other studies, Eisenhardt (1985, 1988) exam-ined the choice between commission (outcome-based) and salary (behavior-based) compensa-tion of salespeople in retaiUng. The originalstudy (1985) included only agency variables,while a later study (1988} added additionalagency variables and institutional theory pre-dictions. The results supported agency theorypredictions that task programmability, informa-tion systems (measured by the span of control),and outcome uncertainty variables (measured

69

by number of competitors and failure rates) sig-nificantly predict the salary versus commissionchoice. Institutional variables were significantas well.

Conlon and Parks (1988) repUcated and ex-tended Eisenhardt's work in a laboratory set-ting. They used a multiperiod design to test bothagency and institutional predictions. Consistentwith agency theory (Harris & Raviv, 1978), theyfound that information systems (manipulated bywhether or not the principal could monitor theagent's behavior) were negatively related toperformance-contingent (outcome-based) pay.They also found support for the institutional pre-dictions.

Finally, Eccles (1985) used agency theory todevelop a framework for understanding transferpricing. Using interviews with 150 executives in13 large corporations, he developed a frame-work based on notions of agency and faimess toprescribe the conditions under which varioussourclng and transfer pricing altematives areboth efficient and equitable. Prominent in hisframework is the link between decentralization(arguably a measure of task programmability)and the choice between cost (behavior-basedcontract) and market (outcome-based contract)transfer pricing mechanisms.

In summary, there is support for the principal-agent hypotheses linking contract form with (a)information systems (Conlon & Parks, 1988; Ec-cles, 1985; Eisenhardt, 1985), (b) outcome uncer-tainty (Eisenhardt, 1985), (c) outcome measur-ability (Anderson, 1985; Eisenhardt, 1985), (d)time (Conlon & Parks, 1988), and (e) task pro-grammability (Eccles, 1985; Eisenhardt, 1985).Moreover, this support rests on research using avariety of methods including questionnaires,secondary sources, laboratory experiments,and interviews.

Recommendations for AgencyTheory Research

As argued above, agency theory makes con-tributions to organization theory, is testable, and

has empirical support. Overall, it seems reason-able to urge the adoption of an agency theoryperspective when investigating the many prob-lems that have a principal-agent structure. Fivespecific recommendations are outUned belowfor using agency theory in organizational re-search.

Focus on Information Systems, OutcomeUncertainty, and Risk

McGrath, Martin, and Kukla (1981) arguedthat research is a knowledge accrual process.Using this accrual criterion, next steps foragency theory research are clear: Researchersshould focus on information systems, outcomeuncertainty, and risk. These agency variablesmake the most unique contribution to organiza-tional research, yet they have received little em-pirical attention (Table 3). It is important that re-searchers place emphasis on these variables inorder to advance agency theory and to providenew concepts in the study of familiar topics suchas impression management, innovation, verti-cal integration, compensation, strategic alli-ances, and board relationships.

Studying risk and outcome uncertainty is par-ticularly opportune because of recent advancesin measuring risk preferences. By relying on theworks of Kahneman and Tversky (1979), Mac-Crimmon and Wehrung (1986), and March andShapira (1987), the organizational researchercan measure risk preference more easily andrealistically. These techniques include directmeasures of risk preference such as lotteries andindirect measures using demographic charac-teristics such as age and wealth and payoffcharacteristics such as gain versus loss. (SeeMarch and Shapira, 1987, for a review.)

Key on Theory-Relevant Contexts

Organizational theory usually is explored insettings in which the theory appears to havegreatest relevance. For example, institutionaland resource dependence theories were devel-oped primarily in large, public bureaucracies inwhich efficiency may not have been a pressing

70

concem. The recommendation here is to takethe same approach with agency theory: Key ontheory-relevant contexts.

Agency theory is most relevant in situations inwhich contracting problems are difficult. Theseinclude situations in which there is (a) substan-tial goal conflict between principals and agents,such that agent opporiunism is likely (e.g., own-ers and managers, managers and profession-als, suppliers and buyers); (b) sufficient outcomeuncertainty to trigger the risk implications of thetheory (e.g., new product innovation, youngand small firms, recently deregulated indus-tries); and (c) unprogrammed or team-orientedjobs in which evaluation of behaviors is difficult.By emphasizing these contexts, researchers canuse agency theory where it can provide the mostleverage and where it can be most rigorouslytested. Topics such as innovation and settingssuch as technology-based firms are pariicularlyattractive because they combine goal conflictbetween professionals and managers, risk, andjobs in which performance evaluation Is diffi-cult.

Expand to Richer Contexts

Perrow (1986) and others have criticized agencytheory for being excessively narrow and havingfew testable implications. Although these criti-cisms may be extreme, they do suggest that re-search should be undertaken in new areas.Thus, the recommendation is io expand to aricher and more complex range of contexts.

Two areas are particularly appropriate. Oneis to apply the agency structure to organiza-tional behavior topics that relate to informationasymmetry (or deception) in cooperative situa-tions. Examples of such topics are impressionmanagement (Gardner & Martinko, 1988), lyingand other forms of secrecy (Sitkin, 1987), andblame (Leatherwood & Conlon, 1987). Agencytheory might contribute an overall framework inwhich to place these various forms of self-interest, leading to a better understanding of

[when such behaviors will be likely and whenPthey will be effective.

The second area is expansion beyond thepure forms of behavior and outcome contracts asdescribed in this article to a broader range of con-tract altematives. Most research (e.g., Anderson,1985; Eisenhardt, 1985, 1988) treats contracts as adichotomy: behavior versus outcome. However,contracts can vary on a continuum between be-havior and outcome contracts, Also, current re-search focuses on a single reward, neglectingmany situations In which there are multiple re-wards, differing by time frame and contract ba-sis. For example, upper level managers usuallyare compensated through multiple rewardssuch as promotions, stock options, and salary.Both multiple and mixed rewards (behavior andoutcome) present empirical difficulties, but theyalso mirror real life. The richness and complex-ity of agency theory would be enhanced if re-searchers would consider this broader spectrumof possible contracts.

Use Multiple Theories

A recent article by Hirsch et al. (1987) elo-quently compared economics with sociology.They argued that economics is dominated by asingle paradigm, price theory, and a singleview of human nature, self-interest. In contrast,the authors maintained that a strength of orga-nizational research is its polyglot of theories thatyields a more realistic view of organizations.

Consistent with the Hirsch et al. arguments,the recommendation here is fo use agency the-ory with complementary theories. Agency the-ory presents a partial view of the world that,although it is valid, also ignores a good bit of thecomplexity of organizations. Additional per-spectives can help to capture the greater com-plexity.

This point is demonstrated by many of the em-pirical studies reviewed above. For example,the Singh and Harianto (in press) and Kosnik(1987) studies support agency theory hypothe-ses, but they also use the complementary per-spectives of hegemony and managerialism.These perspectives emphasize the power and po-litical aspects of golden parachutes and green-

71

mail, respectively. Similarly, the studies byEisenhardt (1988) and Conlon and Parks (1988)combine institutional and agency theories. Theinstitutional emphasis on tradition complementsthe efficiency emphasis of agency theory, andthe result is a better understanding of compen-sation. Other examples include Anderson (1985),who coupled agency and transaction cost, andEccles (1985), who combined agency with equitytheory.

Look Beyond Economics

The final recommendation is that organiza-tional researchers should look beyond the eco-nomics literature. The advantages of economicsare careful development of assumptions andlogical propositions (Hirsch et al., 1987). How-ever, much of this careful theoretical develop-ment has already been accomplished for agencytheory. For organizational researchers, the pay-off now is in empirical research, where organi-zational researchers have comparative advan-tage (Hirsch et al., 1987). To rely too heavily oneconomics with its restrictive assumptions suchas efficient markets and its single-perspective

style is to risk doing second-rate economics with-out contributing first-rate organizational re-search. Therefore, although it is appropriate tomonitor developments in economics, it is moreuseful to treat economics as an adjunct to moremainstream empirical work by organizationalscholars.

Conclusion

This paper began with two extreme positionson agency theory—one arguing that agencytheory is revolutionary and a powerful founda-tion (Jensen, 1983) and the other arguing that thetheory addresses no clear problem, is narrow,lacks testable implications, and is dangerous(Perrow, 1986). A more vaUd perspective lies inthe middle. Agency theory provides a unique,reaUstic, and empirically testable perspectiveon problems of cooperative effort. The intent oithis paper is to clarify some of the confusion sur-rounding agency theory and to lead organiza-tional scholars to use agency theory in theirstudy of the broad range of principal-agent is-sues facing firms.

References

Anderaon, E. (1985) The salesperson as outside agent of em-ployee: A transaction cost analysis. Marketing Science. 4,234-254.

Amihud, Y., 8f Lev. B. (1981) Risk reduction as a managerialmotive for conglomerate mergers. Bell Journal of Econom-ics. 12, 605-616.

Argawal, A., & Mandelker, G. (1987) Managerial incentivesand corporate investment and financing decisions. Jour-nal of Finance. 42, 823-837.

Arrow, K. (1971) Essays in the fheory of risk bearing. Chi-cago: Markham.

Bernard, C. (1938) The functions of the executive. Cam-bridge, MA; Harvard University Press.

Barney, J. {1988) Agency theory, employee stock ownershipand a firm's cost of equity capita!. Unpublished workingpaper, Texas A&M University, College Station.

Bamey, I., & Ouchi. W. (Eds.) (1986) Organizafionai econom-ics. San Francisco: jossey-Bass.

Basu, A., Lai, R., Srinivasan, V., & Staelin, R. (1985) Sales-force compensation plans: An agency theoretic perspec-tive. Marketing Science, 4, 267-291.

Berle, A., & Means, G. (1932) The modern corporation andprivate property. New York: Macniillan.

Bolton, M. (1988) Organizationai miming: When do lateadopters of organizational innovations outperform pio-neers? Paper presented at the meeting oi the Academy ofManagement, Anaheim, CA.

Burt, R. (1979) A structural theory oi interlocking corporatedirectorates. SociaJ Networks. 1, 415-435.

Capitalism in the making. (1984, April 30) Time, p. 62.

Chandler, A. (1962) Strcrtegy and structure. New York: Dou-bleday.

Canlon, E., & Parks, I. (1988) The effects of monitoring andtradition on compensation arrangements: An experimenton principal/agent dyads. In F. Hoy (Ed.), Best papers pro-

72

ceedings (pp. 191-195). Anaheim, CA: Academy of Man-agement.

Cyert, R.. & March, J. (1963) A behaviorai (heoiy of the firm.Englewood Cliffs, NI: Prentice-Hall.

Delong, D., Forsythe, R., & Uecker, W. (1985) Ripoffs, lemonsand reputation formation in agency relationships: A lab-oratory market study. Journal of Finance. 50, 809-820.

Demski, J. (1980) A simpJe case of indeterminate financialreporting. Working paper, Stanford University.

Demski, J.. & Feitham, G. (1978) Economic mcentives in bud-getary control systems. Accounting fleview, 53, 336-359.

Dombusch, S., & Scott, W. R. (1974) Evaiudtion and the ex-ercise of authority. San Francisco: Jossey-Bass.

f Eccles, R. (1985) Transfer pricing as a problem oi agency. InI. Pratt & R. Zeckhauser (Eds.), Principals and agents: Thestructure of business (pp. 151-186). Boston: Harvard Busi-ness School Press.

JEisenhardt, K. (1985) Control: Organizational and economicapproaches. Management Science. 31, 134-149.

1 Eisenhardt, K. (1988) Agency and institutional explanationsof compensation in retail sales. Academy of ManagementJournal. 31, 488-511.

iFama, E. (1980) Agency problems and the theory of the firm.Journal of Political Economy. 88, 288-307.

[Fama, E., & lensen, M. (1983) Separation of ownership andcontrol. Journal of Law and Economics. 26, 301-325.

uraith, ]. (1973) I3esigning compJez organizcrtions. Read-ing, MA: Addison-Wesley.

ardner, W., & Martinko, M. (1988) Impression manage-ment: An observational study linking audience character-istics with verbal self-presentations. Academy of Manage-ment Journal. 31, 42-65.

ch, J., & Weiss, A. (1981) Self-selection in the labor mar-ket. American Economic fieview, 71, 275-284.

^arris, M., & Raviv, A. (1978) Some results on incentive con-tracts with application to education and employment,health insurance, and law enforcement. American Eco-nomic Review. 68, 20-30.

larris, M., & Raviv, A. (1979) Optimal incentive contractswith imperiect information. yournaJ of Economic Theory,20, 231-259.

::h. P., & Friedman, R. (1986) Collaboration or paradigmshift? Economic vs. behavioral thinking about policy? In I.Pearce & R. Robinson (Eds.), Best papers proceedings (pp.31-35). Chicago: Academy of Management.

:h. P., Michaels, S., 8r Friedman, R. (1987) "Dirty hands"versus "clean models": Is sociology in danger of beingseduced by economics? Theory and Society, 317-336.

Holmstrom, B. (1979) Moral hazard and observability. BellJournal of Economics. 10, 74-91.

Jensen, M. (1983) Organization theory and methodology. Ac-counting fleview, 56, 319-338.

Jer:sen, M. (1984) Takeovers: Folklore and science. HarvardBusiness Review, 62(6), 109-121.

Jensen, M., & Meckling. W. (1976) Theory o( the firm: Man-agerial behavior, agency costs, and ownership structure.Journal of Financial Economics. 3, 305-360.

lensen, M., & Roeback, R. (1983) The market for corporatecontrol: Empirical evidence. Journal of Financial Econom-ics. II. 5 -50 .

Kahneman, D., & Tversky, A. (1979) Prospect theory: Ananalysis of decisions under risk. Economefrica, 47, 263-291.

Kosnik, R. (1987) Greenmail; A study in board performancein corporate governance. Adminisfrative Science Ouar-terly. 32, 163-185.

Lambert, R. (1983) Long-term contracts and moral hazard.Belljournal of Economics. 14, 441-452.

Lawrence, P., St Lorsch, ]. (1967) Organization and environ-ment. Boston: Division of Research, Harvard BusinessSchool

Leatherwood, M., & Conlon, E. (1987) Diffusibility of blame:Effects on persistence in a project. Academy of Manage-ment Journal 30, 836-848.

MacCrimmon. K., & Wehrung, D. (1986) Tafcing risks: Themanagement of uncertainty. New York: Free Press.

March, ]. (1962) The business firm as a political coalition.Journal of Poliiics. 24, 662-678.

March, }.. & Shapira, Z. (1987) Managerial perspectives onrisk and risk taking. Management Science, 33, 1404-1418.

March, )., & Simon, H. (1958) Organizations. New York:Wiley.

McGrath, I. Martin, I., & Kukla, R. (1982) Judgment calls inresearch. Beverly Hills, CA; Sage.

Meet Mike Jensen, the professor of merger mania. (1988,February 8) Business Week. pp. 66-67.

Mitnick, B. (1966) The theory of agency and organizationalanalysis. Unpublished working paper. University of Pitts-burgh.

Ouchi, W. (1979) A conceptual framework ior the design oforganizational control mechanisms. Management Sci-ence, 25, 833-848.

Pearoe, J., Stevenson, W., & Perry, J. (1985) Managerial com-pensation based on organizational performance: A timeseries analysis of the effects of merit pay. Aoademy ofManagement Journal. 28, 261-278.

73

Perrow, C. (1986) Complex organizations. New York: Ran-dom House.

Pettigrew, A. (1973) The politics of organizafionai decisionmaking. London: Tavistock.

Pfeffer, J. (1981) Power in organizations. Marshfield, MA: Pitt-man.

Pfeffer, J., & Salancik. G. (1974) Organizational decisionmaking as a political process: The case of a universitybudget. Administrative Science Quarferiy, 19, 135-151.

Ross, S. (1973) The economic theory of agency; The princi-pal's problem. American Economic Review. 63. 134-139.

Shaveli, S. (1979) Risksharing and incentives in the principaland agent relationship. Bell/ournai of Economics, 10, 53-73.

Singh, H., & Harianto, F. (in press) Management-board re-lationships, takeover risk and the adoption of golden para-chutes: An empirical investigation. Academy of Manage-ment Journal

Sitkin, S. (1987) Secrecy in organizations: The limits of legit-imate information control Working paper. University ofTexas, Austin.

Spence, A. M. (1975)The economics of internal organization;An introduction. BeJJ/ournaJ of Economics, 6, 163-172.

Spence, A. M., 8t Zeckhauser, R. (1971) Insurance, informa-tion, and individual action. American Economic fleview,61, 380-387.

Thompson, ]. (1967) Organizations in action. New York; Mc-Graw-Hill.

Ungson, G., & Steers, R. (1984) Motivation and poUtics inexecutive compensation. Academy of Management Re-view. 9, 313-323.

Walker, G.. & Weber, D. (1984) A transaction cost approachto make-or-buy decisions. Administrative Science Quar-terly. 29. 373-391.

Walking, R, & Long. M. (1984) Agency theory, managerialwelfare, and takeover bid resistance. The Rand Journal olEconomics, 15, 54-68.

White, H. (1985) Agency as control In J. Pratt & R. Zeck-hauser (Eds.), Principals and agents: The structure of busi-ness (pp. 187-214). Boston; Harvard Business SchoolPress.

Williamson, O. (1975) Markets and hierarchies: Analysis andantitrust implications. New York; Free Press.

Wilson, R. (1968) On the theory of syndicates. Econometrico,36. 119-132.

Wolfson. M. (1985) Empirical evidence of incentive problemsand their mitigation in oil and gas shelter programs. In I.Pratt & R. Zeckhauser (Eds.), Principais and agents.- Thestructure of business (pp. 101-126). Boston; Harvard Busi-ness School Press.

Zenger, T. (1988) Agency sorting, agent so/utions and disecon-omies of scale: An empiricai investigation of empioymen:contracts in high technology R&D. Paper presented at themeeting of the Academy of Management, Anaheim, CA.

Kathleen M. Eisenhardt (Ph.D.. Stanford University) isAssistant Professor at Stanford University. Correspon-dence can be sent io her at ihe Department of Indus-trial Engineering and Engineering Management. 346Terman Building. Stanford University. Stanford. CA94305.

The author thanks Paul Adler. Michele Boiton. PhilipBromiiey. Jim Hodder, William Ouchi. Gerald Salan-cik. Kaye Schoonhoven. and Robert Sutton for theircomments and suggestions.

74