principal-agent final paper

7/31/2019 Principal-Agent Final Paper

1/74

Far East University

Angelo Carlo M. Rosario Graduate School of International Development

Major in Business Administration


2/74

The Principal-Agent Theory

Angelo Carlo Rosario

Business Ethics

Spring Semester 2012

Dr. Orapin Sopchokchai

June 06, 2012


3/74

Table of ContentsChapter 1

Executive Summary.....1Introduction..2

What is Agency Relationship? 3Agency Theory.4

Critisms of Agency Theory..4

Importance of Agency Theory.6

Overview......7

Chapter 2

The Principal-Agent Theory8

Origins of Principal-Agent Theory....10Development of the Principal-Agent Theory.16

Agency Relationships18

Types of Agency Relationships.19Levels of Agency Relationships20

Assumptions about Agency Relationship..22

Roles f Agency Theory..24

Corporate Considerations...32

Chapter 3

Principal-Agent Problem...33Agency Problems...33

Problem of Adverse Selection.......34

Agency Cost and Variables38Government Inefficiency...41

Maximizing Utility41

Distorted Bonus Bonanza..42

Corporate Governance Failure...42

Chapter 4

What ethical Issues can arise with Agency Theory? 44Agency and Ethics.44

Conflict of Interest.45

Moral Hazard.45

Tacit Information...46Trust...47

Logical Inconsistency48


4/74

Chapter 5

Summary, Theoretical Implications and Future Research Directions...50

Conclusion.52References....54

List of Tables and Figures

Table 1.140


5/74

Executive Summary

This paper attempts to provide an overview of the major literatures, whichhas developed

in the area of the principal-agent theory. This research is divided into foursections. First, the introduction and overview of the topic with arguments

pertaining to this theory, arguments and the importance of the said concept.Second is the discussion about Principal agency theory, its origins anddevelopment focusing in different aspects of this theory. Third, are the

discussions about the problems of principals and agents. Fourth, the issueabout principals and the agents, which is Ethics, is addressed. Last, asummary and lessons learned about the basic tenets of this paper and

theoretical implications are considered and future research directions arerecommended.


6/74

1.1 Introduction

The principal-agent-theory is a theoretical approach within the social sciences and

delivers a model to explain the behavior of actors within a hierarchical order as well as

general assumptions on explicit as well as implicit treaty formulations. this theory acts onthe assumption that rationally thinking economic actors are constrained within theirdecision making. One possible reason could be an asymmetric distribution of

information. In the model, the principal entrusts his agent with a task. Both of them actaccording to their own interest, which could lead to conflict situation since they could

pursue different aims. But the agent is generally considered to be more risk-averse thanthe principal.

Principal Agent Theory refers to the arrangement that exists when one person or entity(called the agent) acts on behalf of another (called the principal). For example,

shareholders of a company (principals) elect management (agents) to act on their behalf,and investors (principals) choose fund managers (agents) to manage their assets. This

arrangement works well when the agent is an expert at making the necessary decisions,but doesn't work well when the interests of the principal and agent differ substantially. In

general, a contract is used to specify the terms of a principal-agent relationship. (Stevens2002),

The principal expects his agent to completely engage himself in fulfilling his task by

pursuing the principal's aims and not his own. But the principal is limited in recognizingthe engagement as well as the qualities of his agent and is only able to see the final result.

Therefore, the agent possesses an advantage considering information, since he is morecapable of judging the success of his efforts. The agent is now able to utilize this

informational asymmetry in order to follow his own purposes by acting to the

disadvantage of the principal. This theory also underlies most studies of organizationalcontrol in accounting and economics is principal-agent theory (See reviews by Baiman1982, Pratt and Zeckhauser 1985, Eisenhardt 1989, Baiman 1990, and Mitnick 1992).

Traditional principal-agent models, however, assumes that individuals are self-interested,opportunistic, and motivated solely by wealth andleisure (Luft 1997). In particular,

individuals are indifferent to the utility of others or abstract values such as honesty orduty (Koford and Penno 1992).

Based on this behavioral assumption, principal-agent models prescribe complex incentive

schemes and costly monitoring to control opportunistic behavior within the organization.Employment contracts found in practice, however, are simple and incomplete and the

level of monitoring is far less than that required by the theory (Stiglitz 1991). This has ledeconomists to search for ways to enhance the external validity of the theory (Akerlof

1984; Simon 1991; Besley and Ghatak 2003)

Researchers have long asserted that incorporating ethics could enhance both the internaland external validity of agency theory. Noreen (1988) argued that ethical behavior makes

the formation of organizations possible. Koford and Penno (1992) asserted thatmost people have attitudes toward telling the truth and providing fair amounts of effort,


7/74

and agency models neglect a significant element of reality by failing to incorporate suchattitudes. Luft(1997) argued that integrating ethics within agency theory could yield new

insights regarding transaction costs within the organization and the effects of accountinginformation.

Results from recent experimental studies also suggest that concerns for ethics are amajor motivator. For example, researchers have documented that human subjectssacrifice earnings to honestly report their production potential (Evans et al. 2001), reduce

budgetary slack (Stevens 2002), reduce shirking (Schatzberg and Stevens 2005), andreciprocate for acts of kindness (Hannan 2005). Of particular relevance to agency theory,

Stevens (2002) demonstrates that concerns for ethics can affect behavior independent ofconcerns for reputation. Despite the growing evidence that integrating ethics with

economic theory would improve internal and external validity, we are not aware of anypublished study that integrates ethics with traditional principal-agent theory.

To address this void in the literature, we study the feasibility and desirability of this

integrative approach. Specifically, we introduce a principal-agent model with ethics andinvestigate the properties of an ethics-based solution to the moral hazard problem In the

traditional principal-agent framework, the principal offers the agent a wagefor performing a productive effort, and the agent accepts the offer as long it adequately

compensates him for his effort. After accepting the offer, however, the agent prefers toshirk his moral obligation to the principal because he is opportunistic and cares only

about wealth and leisure. When effort is unobservable, the agents potential moral failuregives rise to a moral hazard problem for the principal. Traditional solutions to this moral

hazard problem involve financial incentives or external monitoring. We examine a thirdsolution that involves internal or self-monitoring. In particular, we examine a solution

that relies on the likelihood that the agent will suffer at least some disutility for shirkinghis moral obligation to the principal. This disutility arises because shirking violates the

ethical norm that valid-agreements-should-be-kept.(Mitnick 1992). It differs by agentand may still be zero as in the traditional case

1.2 What is an agency relationship?

An agency relationship arises when one or more principals (e.g. an owner) engage

another person as their agent (or steward) to perform a service on their behalf.Performance of this service results in the delegation of some decision-making authority to

the agent. This delegation of responsibility by the principal and the resulting division oflabor are helpful in promoting an efficient and productive economy. However, such

delegation also means that the principal needs to place trust in an agent to act in theprincipals best interests. What happens when concerns arise over the motives of agents

and cause principals to question the trust they place in them?


8/74

1.3 Agency Theory

Agency Theory refers to analyses associate with the Principal-Agent Relationship which

occurs whenever one person acts in the interests of another. Many situations create a

principal-agent relationship between two people. Explicit relationships include those

situations where one person acts in the interests of another through contractualagreements. For example, when owners of a corporation hire a manager to run thecompany using his/her expertise and experience, a formal contract is created where the

managers act in the interests of the owners in exchange for compensation such as asalary, stocks, and even perquisites (Mitnick 1992).

A simple agency model suggests that, as a result of information asymmetries and self -

interest, principals lack reasons to trust their agents and will seek to resolve theseconcerns by putting in place mechanisms to align the interests of agents with principals

and to reduce the scope for information asymmetries and opportunistic behavior. Motivesof agents and information asymmetries Agents are likely to have different motives to

principals.

They may be influenced by factors such as financial rewards, labor market opportunities,and relationships with other parties that are not directly relevant to principals. This can,

for example, result in a tendency for agents to be more optimistic about the economicperformance of an entity or their performance under a contract than the reality would

suggest. Agents may also be more risk averse than principals. As a result of thesediffering interests, agents may have an incentive to bias information flows. Principals

may also express concerns about in formation asymmetries where agents are inpossession of information to which principals do not have access.

Some principal-agent relationships do not operate formally but exist as though there isagency. For example, employees and managers in a corporation do not have a formalagency relationship. This is because the managers do not themselves compensate the

employees for working on behalf of the managers. However, at some point in theirrelationship, the managers must rely on the employees without monitoring them all the

time. This relationship can still be defined and governed by theories of agency. (Evans etal. 2001)

1.4 Criticisms of Agency Theory

Agency theory remains somewhat controversial, particularly in the literature on financialethics. For example, Horrigan claims that agency theory "raises the ethical danger of

creating a very contentious, litigious view of financial relationships, pitting agents againstprincipals and principals against principals as perpetual adversaries."That the business

world is a litigious place very few would deny. It has been so for some time for longer,even, than Jensen and Meckling's article has been in print. Horrigan's view that the

business world is an adversarial place is valid also, but this also hardly originated withagency theory. The world of business has competition at its very core. Competition exists

not only among firms, but within firms, as employees compete for recognition,


9/74

promotions, and salary increases. Agency theory acknowledges this world, but did notcreate it Dobson claims to be quoting Jensen and Meckling when he says that agency

theory is based on the view that management's thirst for perquisites at the expense offirm value is as inviolable as 'a world in which iron ore is a scarce commodity.' In fact,

Dobson quotes this passage out of context. The passage merely refers to the pointlessness

of comparing outcomes in a world with agency costs to outcomes where the presence ofagency costs are denied, and saying that the latter is preferable. The full passage is below:

In conclusion, finding that agency costs are non-zero and concluding thereforethat the agency relationship is non-optimal, wasteful, or inefficient is equivalent in everysense to comparing a world in which iron ore is a scarce commodity (and thereforecostly) to a world in which it is freely available at zero resource cost, and concludingthat the first world is 'non-optimal'"

Brennan objects to the "cynicism" and "harshness" of agency theory. He says that "theprerequisite consumption model of Jensen and Mecklingrests on the assumption that a

manager (agent) will steal what he does not own, so that it is probably more efficient togive it to him at the outset rather than put him to the trouble of stealing it." Brennan goes

on to state that such agents should probably be replaced rather than tolerated within theorganization. Brennan's recommendation about replacement is probably sound. It is

(perhaps) made in tongue-in-cheek fashion, exaggerated (in Brennan's mind) for effect.However, it may be a denial of reality. For example, retailers traditionally have suffered

more shoplifting losses from their employees than from the public. Those who steal arefired when caught, but it doesn't seem to slow the tide of losses.

It seems that many of the criticisms of agency theory rest on an implicit assumption of

the perfectibility of man. The critics imply that if honesty and virtue were only givengreater emphasis in our moral discussion, then people will behave in a more honest,

virtuous manner. For example, Horrigan says that "Positive ideas inevitably becomenormative ideas when they are promulgated in decision oriented subjects, such as

financial management." Dobson agrees, saying that "By assuming unbridled self-interest,financial economics promotes unbridled self-interest." However, evidence of the innate

sinfulness of mankind is all around us. The Scriptures admonish Christians not to becometoo comfortable with their moral righteousness. The Apostle Paul, in the Book of

Romans, compiles quotations from several Old Testament passages when he writes:

As it is written: There is no one righteous, not even one; there is no one whounderstands, no one who seeks God. All have turned away, they have together becomeworthless; there is no one who does good, not even one. Their throats are open graves;their tongues practice deceit. The poison of vipers is on their lips. Their mouths are fullof cursing and bitterness. Their feet are swift to shed blood.; ruin and misery mark theirways, and the way of peace they do not know.There is no fear of God before their eyes.

Paul sums it up by saying "For all have sinned and fall short of the glory of God.


10/74

If any theory or set of ideas is going to help develop the moral sensibilities of managers,it must start with a view of the (business) world in which is accurate in every respect. We

cannot build a moral and virtuous world by assuming that immorality does not exist, norby assuming that mankind is perfectible or is naturally evolving into such a state. The

purposes of reform are best accomplished by starting with a realistic view of the world

and of the nature of mankind. Agency theory does this, but receives criticism for it.Agency theory is grounded in the view that mankind is not perfect, nor naturally evolvingin that direction in assuming self-interest as the primary motive of mankind, agency

theory makes no claim that self-interest is morally preferable. In this sense, it recognizesour fallen state.

1.5 The importance of the Agency Theory (Principal- Agent)

The concepts of agency theory has become a lot more important now than ever before,

especially with the corporate collapses of major multinational companies such as Enronin 2001, WorldCom in 2001, Lehman brothers 2008, and the nationalization of many

financial institutions in recent months as a result of the global financial crisis. Lots ofcompanies have been nationalized in recent times than ever before across all developed

economies, such companies include Washington mutual, Bradford and Bingley, Northernrock and the like.

Agency theory or agency relationship is the theory which looks at the relationship

between the owners of the company in the form of shareholders (equity investors) andthose have been given responsibility to take charge of the management of the company in

the form of directors. Agency theory is one of the key concepts underlying theimportance of corporate governance, which has taken prominent role in business

activities in the last few decades. (Armah, 2002)

It has its roots in almost every aspect of business activities and plays a very significant

role in decision-making by directors (both executive and non-executive directors. And wemust know and understand how Principal and Agents act upon these.

Agency theory has been considered especially valuable in re-establishing the importance

of incentives, interests and information in organizational thinking. It assumes thatwhether we like it or not, much of organizational life is partly based on peoples self

interest, opportunism and goal conflicts. In addition, the theory has drawn attention to theissues related to information, and especially the asymmetries of information. An

important contribution of agency theory is its risk implications. Organizations areassumed to have uncertain futures. Agency theory extends the ramifications of outcome

uncertainty to their implications for creating risk. Uncertainty is viewed as risk / rewardtrade-offs with the implication that outcome uncertainty coupled with differences in

willingness to accept risk should influence contracts between the principal and the agents.(Eisenhardt, 1989, pp. 64-65.) Agency theory draws attention, above all, to the theoretical

and practical issues related to control. The theory asserts that there are two.


11/74

The criticisms of Donaldson are also orientated towards Organizational Economics ingeneral. Donaldson discusses agency theory mainly in the context of corporation

management. However, the criticism is so general that it can also be applied to otheragency relationships. Problems for intra- and inter-organizational settings: the problems

of adverse selection and moral hazard. The theory also points to perspectives and

mechanisms for overcoming these problems. Recognizing the problem of adverseselection offers more understanding, for example, as to why the recruitment of an agentmight sometimes fail, and why there is a need to reduce information asymmetries

between the principal and the prospective agent.

The problem of moral hazard, on the other hand, has increased the understanding of whyorganizations or individuals invest in monitoring mechanisms, and why piece-rate

salaries are sometimes used. Additionally, agency variables have been able to clarify thebenefits and shortcomings of different control procedures in a certain context. In general,

it seems that agency theory has made its greatest contributions in understanding thenature of the goal conflicts that can arise between principals and agents, informational

asymmetries and the potential problems that result from the different forms of agentopportunism, and the governance structures that evolved to contain them (cf. Davis et al.,

1997, p. 45). The scholarly interest in agency theory has seemingly not decreased in thelast two decades. On contrary, new fields of applications are being introduced

continuously. Also, the numerous attempts to develop the theory further can also beinterpreted as confirmation of the general perspective and problems that agency theory

attempts to address. Although the descriptive, explanatory and predictive qualities ofagency theory still seem to have remained context-bound and debated, the theory has

offered alternative views for understanding the dynamics of different types of agencyrelationships. As a framework it has provided interesting new ways of thinking (Koelble,

1996, p. 261), it has had considerable value in helping to sort out and clarify relationshipsof power (Laffin, 1997, p. 56), and contributed understanding of existing governance

structures (Ferris & Graddy, 1998, p. 228)

1.6 Overview

In this paper we would discuss some arguments and problems about the agency theory,the impact of this theory to each sector maybe in business, economy, ethical standards

and the like, the term Principal and Agent which is indeed used in widespread inartificial intelligence specially in complex systems, and regarded as one of the most

important foundations of this discipline. This paper will help us understand the role andthe problems in a easier sense, magnifying some arguments and of past researchers andparticularly discussing some of the actors role in different settings,


12/74

Chapter 2

2.1 The Principal-Agent Theory

The Principal agent theory is a theory explaining the relationship between principals,

such as a shareholders, and agents, such as a company's executives. In this relationshipthe principal delegates or hires an agent to perform work. The theory attempts to dealwith two specific problems: first, that the goals of the principal and agent are not in

conflict (agency problem), and second, that the principal and agent reconcile differenttolerances for risk. Agency theory is a concept that explains why behavior or decisions

vary when exhibited by members of a group. Specifically, it describes the relationshipbetween one party, called the principal that delegates work to another, called the agent. It

explains their differences in behavior or decisions by noting that the two parties oftenhave different goals and, independent of their respective goals, may have different

attitudes toward risk.

Principal-Agent theory emerges with the division of labor and exchange. The division oflabor induces the need for delegation and the first historical contracts appear probably in

agri-culture when a landlord (principal) contracts with his tenant (agent). It is then nowonder that Adam Smith encountered incentive problems in his discussion of this theory

The concept of this theory originated from the work of Adolf Augustus Berle and

Gardiner Coit Means, who were discussing the issues of the agent and the principal asearly as 1932. Berle and Means explored the concepts of agency and their applications

toward the development of large corporations. They saw how the interests of the directorsand managers of a given firm differ from those of the owner of the firm, and used the

concepts of agency and principal to explain the origins of those conflicts.

Michael C. Jensen and William Meckling shaped the work of Berle and Means in thecontext of the risk-sharing research popular in the 1960s and '70s to develop agency

theory as a formal concept. Jensen and Meckling formed a school of thought arguing thatcorporations are structured to minimize the costs of getting agents to follow the direction

and interests of the principals

Principal-agent theory arises in a business management context associated with

behavioral studies of employer-contractor or employer-employee interactions but it canbe applied to public and non-profit settings as well. Early work centered on dilemmas of

dealing with incomplete information in insurance industry contracts (Spence andZeckhauser, 1971; Ross, 1973). The theory was soon generalized to dilemmas associated

with contracts in other contexts (Jenson and Meckling, 1976; Harris and Raviv, 1978).Because some research in this area utilizes experiments in small group interaction, there

is a close relation to game theory, as some principal-agent writers make explicit criticismabout this theory


13/74

When you hire a solicitor, however, it is almost impossible for you to monitor her effortand diligence on your behalf. You have not studied law, and much of what the solicitor

does will be a mystery to you.

This latter situation is close to the relationship that exists between shareholders and

managers. The managers have information and expertise that the shareholders do nothave-indeed, that is why they are the managers. The shareholders can observe profits, butthey cannot directly observe the managers' efforts. To complicate matters further, even

when the managers' behavior can be observed, the shareholders do not generally have theexpertise to evaluate it. Everyone can see the firm's revenues, but it takes very detailed

knowledge to estimate how large those revenues could have been if the managers hadacted differently. Boards of directors, who represent the firm's shareholders, can acquire

some of the relevant expertise and monitor managerial behavior, but again this is costly.

Agency theory essentially acknowledges that different parties involved in a givensituation with the same given goal will have different motivations, and that these different

motivations can manifest in divergent ways. It states that there will always be partial goalconflict among parties, efficiency is inseparable from effectiveness, and information will

always be somewhat asymmetric between principal and agent. Agency theory has beensuccessfully applied to a lot of disciplines including accounting, economics, politics,

finance, marketing, and sociology.

Research on principal-agent theory has had several findings. Most notably, an agent is

more likely to adopt the goals of the principal, and thus behave in the interest of theprincipal, when the contract is outcome-based. Also, when the agent is aware of a

mechanism in place that allows the principal to verify the behavior of the agent, he ismore likely to comply with the goals of the principal.

Furthermore, outcome uncertainty has a positive relationship to behavior-based contracts,

while there is a negative relationship to outcome-based contracts. Goal conflict has anegative relationship to behavior-based contracts with a positive relationship toward

outcome-based contracts. Outcome measurability is negatively related to behavior-basedcontracts; there is a positive relationship with respect to outcome-based contracts.

Opponents to agency theory criticize it as being too general and claim that it is pseudo-

scientific. They also claim that its interpretation is subjective and the validity of agencytheory is not testable. The ability to be empirically tested is a necessary component of any

hypothesis.


14/74

2.2 Origins Principal agent Theory

The agency theory paradigm, first formulated in the academic economics literaturein the early 1970s (Ross 1973, Jensen & Meckling 1976) had diffused into the business

schools,

The first scholars to propose, explicitly, that a theory of agency be created, and to

actually begin its creation, were Stephen Ross and Barry Mitnick, independently and

roughly concurrently. Ross is responsible for the origin of the economic theory ofagency, and Mitnick for the institutional theory of agency, though the basic concepts

underlying these approaches are similar. Indeed, the approaches can be seen ascomplementary in their uses of similar concepts under different assumptions. In short,

Ross introduced the study of agency in terms of problems of compensation contracting;agency was seen, in essence, as an incentives problem. Mitnick introduced the now

common insight that institutions form around agency, and evolve to deal with agency, inresponse to the essential imperfection of agency relationships: Behavior never occurs as it

is preferred by the principal because it does not pay to make it perfect. But society createsinstitutions that attend to these imperfections, managing or buffering them, adapting to

them, or becoming chronically distorted by them. Thus, to fully understand agency, weneed both streams -- to see the incentives as well as the institutional structures.

In economic perspective, the problem is one of selecting a compensation system that willproduce behavior by the agent consistent with the principals preferences. Thus the focus

is on the nature of the incentive system and the contracting system that guides thedistribution of those incentives, as well as the conditions of risk and information that

condition the choices of the actors. With his typical elegance, Ross lays out the problemwith great clarity as well as brevity in a paper he delivered at the December 1972

economics meeting and which was published in the AER Proceedings issue in May 1973.He clearly identifies the agency problem as generic in society, not merely as a problem in

the theory of the firm.

This sets his work apart from the existing stream on the theory of the firm (e.g., Baumol1959, Marris 1964, Williamson 1964, Alchian and Demsetz 1972) as well as the more

general formal approaches on decision making under risk or uncertainty, and underdifferent information states (e.g., Arrow 1963, Spence and Zeckhauser 1971, Marshak

and Radner 1972), though it drew much from this work. Ross had recast the problem interms of agency relationships, and clearly identified the key problem and the key

variables. After Rosss paper, scholars would see agency problems and incentivemechanism design issues within agency relationships; the frame of inquiry was

refocused.

For Ross, however, the problems are still within the realm of decisions and sequences of

decisions regarding incentives; the contexts that actually constitute the agencyrelationship are removed from the analysis and are reduced to their contributions of

incentives or contractual constraints or risk/uncertainty conditions to decisions. Workingindependently of Ross, Mitnick followed parallel, if overlapping lines of literature that


15/74

were more institutional in character. Barnard (1938) had observed that in the employmentrelation supervisors and workers observed a zone of discretion or indifference within

which the worker made key choices.

Simon had written about employment as a relation and with March (1958) had developed

an inducements-contributions model that worried explicitly about managerial issues inthe relationship, such as decisions to join and stay in the relationship. Simons workstraddled both economics and public administration, and it was not surprising therefore

that, in political science, Clark and Wilson (1961) developed an incentives model oforganizations. Thus both relational and incentives-based arguments had developed

within, or migrated to, the parallel literature. On the economics side, Williamsonsmanagerial discretion model (1964) recalled, at least superficially, some features of

Barnards old argument about the discretionary zone of managers behaviors, but itintroduced an elegant utility tradeoff in which managers could survive while making

allocation decisions that benefitted themselves as well as giving receivers of profitconsequently less than maximal returns.

Recalling the classic work of Coase (1937), Williamson (1975) then introduced a new

approach to why hierarchies could be superior to market contracting. This transactioncost approach suggested that institutions could form because they were a better means of

dealing with such costs, given other conditions such as the nature of the assets,technology, opportunism, and so on. But although Williamson with others applied this

approach to the employment relation, his transaction costs model (1975) was based on anew view of exchange that is, it existed as a contrast to traditional models of market

exchange and was not offered as a model of control.

His firms, though nominally hierarchical and subject to issues of corporate control by the

owners, were understood functionally more in terms of exchanges transactions rather thanof true hierarchical behavior in which some actors are modeled as acting for others. The

existence of costs of control, however, suggested to Mitnick that a theory of controlcentered on agency not just a theory of exchanges might generate new insights into

common social institutions.

In a 1965 review of theories of the firm, Alchian remarked in analyzing Williamson(1964) that unfortunately for the owners, there are costs of detecting and policing [the

managers] actions. Once these costs are recognized, it is obviously better to avoid someof these costs if the profits saved are less than the costs (Alchian 1965: 35). Mitnick

realized that Alchians observation was a perfectly general one that it was true across thehost of agency relationships, not merely as a characteristic of corporate governance. In

essence, it would be productive to create a vertical theory of control as well as ahorizontal one of exchange. We are then led to focus on the mechanisms, and costs, of

specifying what the agent is to do, as well as the costs of observing and policing him orher. The approach becomes vertically relational, as institutions are created to instruct and

manage agents, and to deal with the inevitable (and sometimes rationally tolerated)imperfections of control.


16/74

Like Ross, Mitnick first presented his work in a conference paper, which was given at aregular panel of the American Political Science Association Meeting in 1973. The APSA

does not publish selected papers in a journal issue, as does the AEA, but it did put panelpapers into microfilmed proceedings that were generally available (as dissertations have

been), from what was Xerox University Microfilms and its corporate heirs.

Both Ross and Mitnick were at the University of Pennsylvania (as was Williamson); Rosswas a young faculty member in economics; Mitnick was a doctoral student in political

science interested in public choice, bureaucracy, and political economy. Mitnickdeveloped his work on agency in the Fall of 1972, writing the long paper that became the

central analytic part of his dissertation and his 1973 APSA paper over a period of monthsup to the 1973 meeting. After his comprehensive examination in early 1973, Prof.

Stephen Elkin asked him what the topic of his dissertation would be. Mitnick describedhis agency thesis, and Elkin said that he should talk to Ross, who had recently presented a

paper on what he called the theory of agency. Mitnick said that he had also come to thatname, having studied the law of agency for his paper and because of simple common

usage of the term agency.

Mitnick did not see Rosss paper until it was published in the May 1973 AERProceedings, however. At that point (perhaps June 1973), he did go to speak with Ross,

and had a brief exchange about the difficulty of creating a formal model of acting for.Ross had a wonderful term for it that stuck in Mitnicks mind the ice cream cone

problem the agent's problem of selecting what the principal wants without knowing theprincipals preferences.

Ross saw the problem as essentially insoluble as a pure choice problem without greaterinformation on those preferences. Mitnick believed that institutions and social

mechanisms exist to guide such behaviors. People make decisions based on things likenorms, information with social origins, and what more recent literature terms cognitive

heuristics or biases. In essence, information can be gathered indirectly or created to solveor remove the ice cream cone problem. Mitnick put a brief summary and critique of

Rosss paper in his own 1973 paper since Rosss paper appeared before the APSAmeeting, but the two works were created quite independently.

Rosss 1973 paper was tightly focused on the formal analysis of the principals problem

of selecting optimal compensation for the agent, and has only the briefest mention of thesocietal contexts and relevance of a theory of agency. He published a second conference

paper in a proceeding in 1974 that focused on the formal principle of similarity. Incontrast, Mitnicks 1973 paper and 1974 dissertation (1974a) presented an extensive

study of many aspects of the theory of agency. It made the case for developing a generaltheory, presenting a detailed set of agency concepts and sorting them in typologies,

identifying types of agency relationships as well as a language for describing agency andfor developing theoretical explanations for behavior in agency.


17/74

For example, Mitnicks work identified the fiduciary norm as a common social norm andnoted how such norms economized on agency costs.

4It also developed a theoretical logic

about agency in general that permitted the generation of theory in a wide variety of socialcontexts. In other words, it actually began developing an institutional theory of agency.

Mitnick presented applications to such social relationships as advisers and clients,

lawyers negotiating with one another, diplomats negotiating with foreign governmentsand one another, the behavioral patterns of legislative representatives, the advocacy ofinterest groups, regulators as agents subject to policing by public observers, regulatory

incentive systems (with a specific application to regulation of power plant siting), and soon. He took the Clark and Wilson (1961) incentive system model and modified it to make

it systematic and applicable to agency relationships.

It was not unusual in the formal work on economic agency to see a focus on normativetheory: For example, researchers asked such questions as, is there an optimal fee schedule

that would align the agent and principal under certain conditions? Approaches of thistype often seek to derive general principles from assumptions about initial conditions,

abstract relationships among variables, and so on, via formal proofs. Having derived theprinciple, some researchers may then search for empirical evidence consistent with the

abstract result.

In contrast, from the beginning Mitnicks institutional approach focused on developingthe core theory logics of agency that made it possible to generate statements about

behavior in the real world, i.e., descriptive theory: How can we explain a series of diversebut commonly observed contexts of agent-principal relationships using the logics of

agency theory? For example, under what conditions are lawyers, acting as advocates inconflict, more or less likely to reach settlement? Why do critics of the performance of

public agencies focus on preventing self-enriching behavior by public employees rather

than on the problem itself, relatively poor outcomes for the public interest? This emphasison descriptive theory continued over the years. For example, Mitnicks more recenttheory of testaments (Mitnick 1996/1998, 1999, 2000), a component of his institutional

agency approach, addresses such questions as, why do selective colleges limit the numberof recommendation letters from applicants? Why do organizations hire imperfectly

trained employees? Why do incumbents have advantages that challengers do not, andwhat factors generate greater advantage than others? All of these apparently diverse

circumstances can be understood by applying the logics of institutional agency theory.

The first regular, non-proceedings journal article on agency as a general theoretical

approach was published by Mitnick (1975b) in Public Choice in the Winter 1975 issue(end of 1975). The widely-cited work by Jensen and Meckling (1976) that proposed an

agency theory of the firm was not published until almost a year later. In 1976 Mitnickpublished another article (1976a) that made use of his agency approach, this time applied

to agency in the public sector, specifically in the context of the public interest and the useof public interest rhetoric in advocacy.

Mitnick also presented several papers on agency

during this period. He was invited by Oliver Williamson to present parts of hisdissertation at Williamson's Organizations Workshop in October 1974 (Mitnick 1974b).

He also presented papers on agency at the American Sociological Association meetings


18/74

in 1975 and 1976 (Mitnick 1975a, 1976b).

The distinguished political scientist Edward C. Banfield served on Mitnick's dissertationcommittee. Banfield passed away a couple of years ago, but two active members of the

committee, the dissertation chair, Russell Hardin, now at New York University, and

Stephen Elkin, now at the University of Maryland, are aware of Mitnick's dissertationwork. Hardin is generally considered one of the leading political philosophers in theworld today, known for his work in public choice, collective action, and trust; Mitnick

was his first doctoral student. Elkin is a distinguished scholar in urban and politicaleconomy who was one of the founders of a prominent organization of political

economists in the U.S., and remains one of its central figures.

Thus, by the time the classic paper by Jensen and Meckling appeared in print in late1976, Ross's economic theory of agency was widely known in economics, and Mitnick's

institutional theory of agency had been published in two articles, used in a third byanother scholar, and been presented at major meetings in three of the social sciences:

economics, political science, and sociology. Although the Jensen and Meckling paper hashad enormous influence in the literature, its occasional citation as the primary originating

paper in agency theory is incorrect. Indeed, it actually originated a variant of an agencytheory of the firm, not agency theory in general.

Of course, the theory of agency did not appear, whole cloth, in the works of Ross andMitnick in 1973. As noted above, key concepts were developed by scholars in economics,

political science, and elsewhere in a variety of streams on the firm, on organizations, andon incentives and information, and were later incorporated into the agency approach. In

addition, agent-principal language was employed in a number of works across the socialsciences well before an explicit theory of agency was proposed. Thus it is important to

both acknowledge the earlier work and recognize that it waited for the primary work byRoss and by Mitnick for the frame of reference to center on agency theory. For example,

in the accounting and control literature, Cooper (1949, 1951) discussed agents inside thefirm; in economics Downs (1957) referred to agents in his economic theory of democracy

and Arrow (1963) referred to agents and to delegation to agents in his discussion ofcharacteristic problems in medical care and the response of institutions to those problems;

in political science, Pitkin (1967) and Tussman (1960) used agent-principal language inworks on political philosophy; and, in sociology, Swanson (1971) described collective

society using such terms. But in none of these works or in any other before Ross andMitnick was there an explicit proposal for, or an actual theory of, agency.

In the early 1970s, agency theory was, of course, unknown in political science and

sociology before Mitnick, and it was not possible to publish articles with this theory logicin journals in these fields. Reviewers wrote that they had never seen anything like it, and

thus it clearly was not sociology or political science. As a result, Mitnick published hiswork on agency in articles and books nominally focusing on other topics. As noted

above, he used an agency approach in part of his paper on the public interest in 1976. Heapplied agency to regulation extensively. Among other applications of agency theory, his

1980 book introduced the study of delegation as the creation of agents in government.


19/74

His edited book on Corporate Political Agency (1993) included both applications ofagency theory and basic theory about agency relationships developed in the context of

corporate political activity.

Agency did not enter political science in a major way until Moe's article in 1984; did not

enter sociology similarly until Shapiro's article in 1987; and did not become prominent inmanagement work until after Eisenhardt's article in 1989. In all cases, Mitnick's work wascited, but it had appeared so many years before that these authors took their primary

direction from the then-popular streams on agency in economics. Moe had been aware ofMitnick's work as long ago as 1978; Shapiro had many of Mitnick's papers and did make

important use of them in her article; Eisenhardt had many of Mitnick's papers, includinghis original papers, but apparently did not consult them and actually cited Mitnick's work

inappropriately as an application in political science rather than as the origination of thefield. Mitnick's original work on agency has indeed been cited (e.g., Cook 1982;

Galaskiewicz 1985; Spulber 1989; Spulber and Besanko 1992; Macey 1992; Watermanand Meier 1998; Krause 1999; there are many more), but it has remained far less known,

especially in economics, than the work that originated with Ross and, later, continuedwith Jensen and Meckling.

Recently, in a major assessment in the Annual Review of Sociology, Shapiro (2005: 4.9,

4.12) noted that a general theory of agency emerged in political science (Mitnick 1973)at the same time that it did in economics (Ross 1973), apparently independently. In a

series of papers spanning at least 25 years, political scientist Barry Mitnick broke themonopoly on agency theory enjoyed by the economics paradigm and offered an

alternative to the assorted baggage that comes with it.

Mitnick continued to publish work using agency as a key theory logic, and to expand the

theory itself.

9

Although he eventually secured a contract from Cambridge UniversityPress to publish his work on agency, the fact of his origination of the theory of agency

seems to have faded in academic memory even as works using agency theory becamecommonplace across the social sciences. Because of the prominence of the Annual

Reviews, Shapiro (2005) may begin a pattern of recalling this work in other responsiblereviews of the literature.

The literatures uncertainty about the origins of agency theory is reflected in the fact that

there is no standard citation to its origin. When work using transaction costs is done, forexample, Oliver Williamson is cited appropriately as its modern originator. When

applications of agency theory are published, however, citations are inconsistent:Sometimes Jensen and Meckling (1976) is cited, although what they originated was an

influential application to the theory of the firm, not the agency approach itself.Frequently, though not always, Ross (1973) is cited, although he deserves to be in every

first footnote because of his critical role in originating economic agency theory. Mitnick(1973, 1975b) is cited far less frequently, although he originated the institutional theory

of agency, including some of agency theorys most basic and familiar concepts andlogics, and was the first to actually make explicit applications of agency theory to social

institutions. Often there are citations to scholars who advanced agency theory or who


20/74

made contributions using agency in the field of the application.

Sometimes there are citations to outstanding scholars who made basic contributions tosome of the core logics that are frequently applied in analyses of agency relationships,

such as formal analyses of the behaviors or effects of risk distributions, incentive

systems, control failures, and the like (e.g., Holmstrm 1979). But there is an extensivehistory of work on such topics. The key argument missing from all earlier work is that ageneral, powerful theory that applies to agency relationships should be developed, and

that such a theory can generate statements about behaviors across the institutional settingsof this type, not merely in one case such as the business firm. And, in surprising number

of cases, there is no citation at all, as if after thirty-plus years the concepts of agency hadpassed into the common language of scholarship.

After Ross (1973), people thought in terms of modeling economic agency relationships,

not just in terms of incentives or compensation systems, which were indeed part ofRosss analysis, and which many scholars had, of course, written about in the past. Part

of the logic of agency costs, including the adaptation of institutions to agency failuresthat may be rationally tolerated or worked around, can now be inferred, for example, in

brief remarks in Alchians (1965) and Alchian and Demsetzs (1972) works on the firm,but there is no recognition of this as a general component of the understanding of how

agency institutions in general are designed and function until Mitnicks work. No one hadoffered that a theory of agency, utilizing, combining, extending, and applying insights

that had appeared across literature in economics, public administration, and politicalscience would be a powerful, new way of viewing a very large class of social phenomena

until Ross and Mitnick proposed just that.

Thus, by the logic of discovery, Mitnick's origination of agency is anchored in his

original creation of the proposal to create a major, integrating theory with this focus, andto actually begin that work, with both presentations and publications of record. For years

after it appeared, there was nothing in the literature that even remotely approached theexplicit and extensive applications that were in his early papers and publications. His

work was by far the first to introduce and develop many of the key institutionalarguments by which we know it today, e.g., the argument that institutions are shaped by

the rational choice of the principal to not create perfect agency, when such creationwould not be worth it. He also was the first to publish a non-proceedings article in a

regular social science journal explicitly on the theory of agency. And the record is clearthat his work was spread even in its early days across three of the social sciences

2.3 Developments of the Principal Agent Theory

The disciplinary origins of agency theory come from economics, more specificallyinformation economics (Eisenhardt, 1989, p. 59). According to Moe (1984, p. 756)

agency theory was initially developed to investigate more general questions ofincomplete information and risk sharing. The works of Spence and Zeckhauser (1971),

Ross (1973), and Arrow (1971), who are usually mentioned as the originators of agencytheory, have been discussed by other authors (see Jensen, 1983, p. 334; Moe, 1984, p.


21/74

756; Eisenhardt, 1989, p. 58). The work of Alchian and Demsetz (1972), which focusedon property rights and addressed issues concerning contracts, shirking and monitoring of

team production, has been influential in the development of agency theory (see Jensen &Meckling, 1976; Fama, 1980).

Since its birth, the development of mainstream agency theoretical research has developedalong two lines, which are usually referred to as the positive theory of agency (a.k.a.positivist stream) and principal- agent (Jensen, 1983, p. 334; Eisenhardt, 1989, p. 59;

Douma & Schreuder, 2002, pp. 109-110). The works of e.g. Ross (1973), Harris & Raviv(1978), Shavell (1979) and Holmstrm (1979) can be considered as early representatives

of principal-agent literature. The work of Jensen &Meckling (1976), and the works ofFama (1980), and Fama and Jensen (1983) were examples of the emerging positive (or

positivist) stream (cf. Jensen, 1983, pp. 334-335). The two streams share a commonunit of analysis, that is, the contract between the principal and the agent, and some of the

common assumptions of agency theory. Jensen (1983, p. 334) argues that bothliteratures address the contracting problem between self-interested maximizing parties

and both use the same agency cost minimizing tautology (although not necessarily statedin that form). According to Eisenhardt (1989, p. 59), the two streams share common

assumptions about people, organizations, and information. Despite their commonassumptions and interests, the streams also differ in many respects. The principal-agent

literature is generally more abstract, mathematical and non-empirically oriented.Principal-agent researchers are concerned about the general theory of the principal-agent

relationship, a theory that can be applied to lawyer-client, landlord-tenant, and employeremployee and other agency relationships.

Characteristic of formal theories, the principal agent stream involves the careful

specification of assumptions, which are followed by logical deduction and mathematicalproof. The main focus is in determining which form of the contract is the optimal one

(Jensen, 1983, p. 334; Eisenhardt, 1989, p. 60). The other stream, the positivist literature,is generally non-mathematical and more empirical in its orientation. Positivist researchers

have focused more on identifying situations in which the principal and the agent arelikely to have conflicting goals and then describing governance mechanisms that limit the

agents self-serving behavior. Positivist researchers have focused more exclusively onintra-organizational principal-agent relationships, especially shareholder-manager

relationships (Jensen, 1983, p. 334; Eisenhardt, 1989, p. 59).

The principal-agent stream has also been developed further in the field of organizationaland management studies. Especially influential have been the works of Kathleen

Eisenhardt (1985; 1988). Eisenhardt developed theoretical and empirical means tomotivate the agent in engaging in desired actions and to reduce the likelihood of shirking.

By following the ideas of Ouchi and Maguire (1975) concerning organizational control,Eisenhardt developed the famous taxonomy between behavior-based and outcome-based

contracts. Further, Eisenhardt also introduced a subset of agency variables to predictwhether the optimal contract is behavior-or outcome-based in a given situation (see

Eisenhardt, 1989). The positive theory of agency seems to have connected to a broaderbody of theoretical work: Organizational Economics (see e.g. Barney & Ouchi, 1986;


22/74

Donaldson, 1990a; Barney & Hesterly, 1996).3 In addition to agency theory,Organizational Economics (OE) is composed of transaction cost economics (see Coase,

1937; Williamson, 1975; 1985) and property rights literature (Alchian & Demsetz, 1972).

Although some other contributions of OE exist (see Barney & Hesterly, 1996), agency

theory and transaction cost economics are clearly its best-known components. As thename implies, OE basically applies different economic models and assumptions to thefield of organization studies. The primary focus of OE has been in explaining the

existence and form of complex organizations (Robins, 1992, p. 524). Organizationaleconomists, such as Barney and Ouchi (1986, xi), denote OE as the study of

organizations and organizational phenomena using concepts taken from contemporaryorganization theory, organizational behavior, and microeconomics. Again, some of the

fierce proponents of OE go further by defining it as an analytical paradigm, a frameworkthat addresses the key determinants of the shape and function of organizations (Hesterly

et al.1990, p. 403, italics in original). Regardless of the questions concerning OEs status,the role of agency theory in organizational economics has been twofold. While it has

shared some of the assumptions of other theories, such as the bounded rationalityassumption from Oliver Williamsons transaction cost theory (Williamson 1975), it has

also contributed to the OE field by introducing concepts such as the agency relationshipand the agency problem, for example. In particular, Jensen & Meckling (1976) have

provided new insights and understanding with regard to agency relationships insideorganizations, by viewing an organization as a nexus of contracts between individuals,

and by introducing the concept of agency costs.

Although the differences between the positive theory of agency and principal-agentstream mentioned above are notable, it is also possible to see these streams as

complementary to each other: Whereas positive theory may identify the behavior of theagent and the various contract alternatives available, the principal-agent stream may

indicate which contract is the most efficient in any given situation (Eisenhardt, 1989,p.60). In addition to economics, the field of political science has been especially active in

applying and developing agency theory.

Political scientist Barry M. Mitnick was probably the first scholar outside the disciplineof economics to recognize the value of the agency framework (see Mitnick, 1975). In the

late 1970s, inspired by the work of e.g. Ross (1973), Susan Rose-Ackerman (1978)distinguished the chain of agency relationships in politics-bureaucracy relationships. In

the early 1980s, the work of political scientist Terry M. Moe also made an importantcontribution in analyzing the theory in different public sector settings. His seminal article

The new economics of organization (1984) has been widely cited by scholars ofdifferent disciplines, including those in the field of economics (see e.g. Eggertsson, 1990)

2.4 Agency relationships

Agency relationships are a general phenomenon. According to Ross (1973, p. 134), The

relationship of agency is one of the oldest and commonest codified modes of socialinteraction and Arrow (1985, p. 37) considers agency relationship as a pervasive fact of


23/74

economic life. Jensen and Meckling (1976) have also highlighted the general nature ofagency relationship by the following arguments: The problem of inducing an agent to

behave as if he were maximizing the principals welfare is quite general. It exists in allorganizations and in all cooperative efforts at every level of management in firms, in

universities, in mutual companies, in cooperatives, in governmental authorities and

bureaus, in unions, and in relationships normally classified as agency relationships suchas is common in the performing arts and the market for real estate. (Jensen & Meckling,1976, p. 309.) In the broadest sense, whenever one party (the principal) depends on the

action of another party (the agent), agency relationship arises (Pratt & Zeckhauser, 1985,p. 3). The basic reason for establishing an agency relationship is usually that the principal

needs a certain task to be performed. The principal acquires the services of the agenttypically because the agent possesses those skills and abilities that are needed for

performing the task. The principal himself may either lack these skills and abilities or heis less effective in performing the tasks than the agent (Petersen, 1993, p. 278).

2.5 Types of Agency Relationship

In most contexts agency relationships are reciprocal (e.g. in the patient doctor

relationship), but they also could be coercive (e.g. in the master slave relationship). Theyoften exist in private corporations and the market environment, but they also can be

found in public bureau and hierarchical environments (Lassar & Kerr, 1996, p. 615;Worsham et al., 1997, p. 430; cf. Moe, 1984, p. 755; Eccles, 1985, p. 159, p. 167).

Traditionally, agency relationships have been considered as contractual relationships (seeJensen & Meckling, 1976; Fama, 1980; Fama & Jensen, 1983).

The contract was the central concept for the early agency theorists because it

distinguished agency theory from classical and neoclassical economics, in which marketforces act as a disciplining mechanism on the owner entrepreneurs who actively manage

firms (Tosi et al., 1997, p. 584). In these contexts, contracts were broadly and vaguelyunderstood to be governance mechanisms ranging in character from formal to informal,

explicit to implicit, and objective to subjective (Barney & Ouchi, 1986, p. 211).Regardless of its character, the main purpose of the contract has been in (a) setting the

tasks for the agent, and (b) introducing the means through which the agent will becompensated (Perrow, 1986, p. 224; Mason & Slack, 2003, p. 40). Thus, the contract in

the agency relationship can be understood to be an instrument enabling different forms ofco-operation and control between the principal and the agent. In fact, some scholars

including Eisenhardt (1989, p. 58) and Bergen et al. (1992, p. 1) have seeminglyinterpreted the contract to be more like a metaphor of an agency relationship, not as a

specific and detailed construct that should be rigorously operationalized.

Agency relationships can range from single-principal-single-agent relationships to morecomplex multiple-principal-single-agent or single-principal- multiple-agent relationships

(cf. Waterman & Meier, 1998, p. 178-180; Mason & Slack, 2003, p. 38). For example,one landlord (the principal) may have only one tenant (the agent), one car repairer (the

agent) may have number of clients (principals) and the employer (the principal) mayemploy number of employees (agents).


24/74

Also, principals and agents may have dual roles in a way that principals can actsimultaneously as some other principals agents, and agents as some other agents

principals. For instance, the chain of representation in parliamentary democracies canillustrate this aspect in following way: the legislators (parliament) are the agents of the

citizenry, but also the principals of the government, which they elect. Likewise, the

government is the agent of the legislators, but also the principal of the nationalbureaucracy. The national bureaucracy, in turn, is the agent of the government, but alsothe principal of the regional or local agencies (Vedung, 1997, p. 107)

2.6 Levels of Agency Relationships

Agency relationships can be distinguished at a range of levels of human co-operation,

from the levels of individuals, groups, or organizations (Smith & Bertozzi, 1998, p. 326).Traditionally, agency theory has been focused on examining the intra-organizational

relationships between individuals (e.g. shareholder-manager, employer-employee) ororganizational groups (shareholders-management teams). In addition to these intra-

organizational analyses, a range of inter-organizational relationships has also beenexamined as agency relationships.

Various scholars outside organizational economics have already used inter-organizational

agency setting in their studies (see e.g. Ferris, 1991; Braun, 1993; Broadbent et al., 1996;Lassar & Kerr, 1996; Ferris & Graddy, 1998). In these studies, the organizational features

of organizations acting as principals and agents have been bypassed and the main focushas been on examining the agency relationship and its behavioral effects on the actions of

the principals and agents.

This implies that neither the purpose nor the analysis of these studies have required exactdefinitions of what constitutes an organization or an organizational principal and that

the assumption of organizational actors as principals and agents has mainly been takenfor granted. Any arguments or analyses on why organizations could or could not be

considered as principals and agents have therefore been lacking.

It seems that early agency theorists have not yet clearly defined how the concept oforganization should be understood. As the theory originates from economics, especially

the versions close to Organizational Economics, applications of the theory have relied onmethodological individualism for their organizational definitions and analyses

(Donaldson 1990a, p. 371).

As the name implies, methodological individualism takes an individualistic perspectiveof organizations. From this perspective, organizations are seen principally as

aggregations of individual preferences and actions, perhaps even only as a nexus ofcontracts (Jensen & Meckling, 1976, p. 310; Pfeffer, 1997, p. 45) or institutional

arrangements that govern the collectivity (Hesterly et al., 1990, p. 405). According tothis perspective, the ultimate participants in organizational activity are individual human

beings, who are and whose actions eventually are the units of all analysis. In thismeaning, organizations are not individuals, they do not have their own preferences,


25/74

motivations or intentions, and they do not exercise choice in the conscious and rationalsense that is attributed only to individual people (Jensen & Meckling, 1976, p. 311;

Jensen, 1983, p. 327).

The organization is essentially a structure designed to harmonize a set of efforts of a

group of people, but it can also be a legal fiction (Jensen & Meckling, 1976, p. 311), ora fictional person who can acquire, hold and dispose of property in its own right(Casson, 1997, p. 79). However, from the perspective of methodological individualism,

the construction named firm or organization itself does not have a real personalityindependent from its individual members People are fundamental first in the sense of

being indivisible decision makers and actors; it is people not organizations whoactually decide, vote or act.

The actions of individuals determine the behavior and performance of organizations.

Furthermore, only the needs, wants, and objectives of individuals have ethicalsignificance. Finally, it is the people who ultimately create and manage organizations,

judge their performance, and redesign or reject them if this performance is foundinadequate. (Milgrom & Roberts, 1992, p. 21.).

On the other hand, understanding the organization as a principal or agent seems not to be

a problem when the organization and its actions are explained through the actions of itsindividual members: Parenthetically, that organizational economics adopts a reductionist

point of view does not mean that this view cannot be applied in the analysis oforganizational phenomena. For some research questions, the theoretical assumption that

firms have a goal or a purpose may fruitfully generate important Insights Even thoughorganizational economists might ultimately attempt to explore the individualistic

underpinnings of these aggregate phenomena, such research does not deny the value of aholistic simplifying assumption in generating the insight in the first place. (Barney, 1990,

pp. 386-387.)

It is possible to suggest that the theorys applicability from the original intra-organizational level to the inter-organizational level becomes a matter of concern only if

the basic assumptions, concepts and conclusions derived from the original setting nolonger hold in the inter-organizational setting. For this reason, the most crucial

requirement for understanding organizations as principals and as agents requires theacknowledgement of some kind of organizational boundaries and set of purposes. This is

because at the individual level, principals and agents are differentiated from otherindividual principals and agents by their observable boundaries and purposes. Therefore,

the acceptability of organizational principals and agents requires that they haveidentifiable legal, economic and / or socio-cultural boundaries and agreed purposes so

that they can be differentiated from other individuals, groups or organizations.Nevertheless, from the perspective of methodological individualism, these purposes (and

the actions based on them) are to be considered only as aggregations of the preferencesand actions of the individual member of an organization.

Assuming the principals and agents to be organizations do not require an understanding


26/74

of them as single-minded and homogenous entities with human characteristics. For thisreason, there is no need to exclude the assumptions concerning intra-organizational

differences of interests, goals and actions of individuals. For example, a human principalmay also have a number of different and competing ideas on how to proceed in a certain

situation with his human agent. Still, he has to choose some of these ideas to put into

practice and has to abandon others. Equally in organizations, individual members of anorganization may have different preferences, but only some of them will be put intopractice as the aggregated will of the whole organization.

2.7 Assumptions about Agency Relationships

The literature on agency theory presents a number of behavioral assumptions concerning

the principal, the agent, and the agency relation Agency ship. It seems that some of theseassumptions depend on different disciplinary and paradigmatic approaches, and they may

even be contradictory.

For instance, the positive agency theoretical literature usually shares the assumption ofbounded rationality with transaction cost theory (Barney &Ouchi, 1986, p. 205;

Eisenhardt, 1989, p. 64; Charreaux, 2002, p. 253). With bounded rationality, humanbehavior is assumed to be intensely rational, but only limitedly so (e.g. Simon, 1957, p.

xxiv). When individuals are bounded rational, they recognize that they cannot possiblyforesee all the things that might concern them.

They understand that communication is imperfect and that understandings are often

flawed, and they know that they are not likely to find the mathematically best solution todifficult problems. They act intentionally though, trying to do the best they can, given the

limitations under which they work (Milgrom & Roberts, 1992, pp. 129-130). On thecontrary, in the more formal, mathematically oriented principal-agent literature,

individuals are assumed to be perfectly rational and to have unlimited computationalabilities. Further, it is assumed that they can anticipate and assess the probability of all

possible future contingencies (Baiman, 1990, p. 342). The principal and the agent areconsidered to be self-interested actors. Additionally, some of the agency theorists

postulate principals and agents as utility maximizers whereas others do not (explicitly)make such an assumption.

The utility maximization assumption is especially important for mathematically oriented

principal agent researchers, because it allows different situations to be modeled andpredicted mathematically in a way that would not be otherwise possible (Hendry, 2005, p.

57). Whether the agents are considered as utility maximizers or not, the assumed self-interest of the agents drives them act opportunistically towards their principal (cf. Barney

& Ouchi, 1986, p. 205). According to Oliver Williamson (1985, p. 47), opportunismrefers to the incomplete or distorted disclosure of information, especially to calculated

efforts to mislead, distort, disguise, obfuscate, or otherwise confuse. It is selfinterestseeking with guile which includes but is scarcely limited to more blatant forms, such as

lying, stealing, and cheating (Williamson, 1985, p. 47). Some agency theorists have alsomade assumptions about the risk preferences of the principal and the agent. The risk


27/74

preference can be than security.

A risk-averse actor prefers security and therefore seeks some guarantee for the attainmentof desirable outcomes. A risk neutral actor, on the other hand, is indifferent to adventure

or security (Bergen et al., 1992, p. 4; see also Pratt, 1964; Arrow, 1971). Usually it is

assumed that the principal is neutral towards the risk and the agent is averse to risk. Therationale is that the agents, who are usually more unable to diversify their employment,are likely to be more risk-averse than the principals, who are usually more capable in

diversifying their investments (Eisenhardt, 1989, pp. 60-61; Bergen et al., 1992, p. 4;Radner, 1992, p. 1406). One important and widely accepted assumption is that agents

differ in their types. The type of agent may refer to such things as whether the agent iscareful versus careless, productive versus unproductive, talented versus untalented,

trustworthy versus untrustworthy, industrious versus lazy, and so forth. In short, it sayssomething about the agents willingness and capacity (i.e. ability) to perform the tasks

agreed on. (Petersen, 1993, p. 278.)

It is also generally assumed, that in addition to the agents actions (a.k.a. effort),environmental factors influence the outcome of the agency relationship (see e.g. Harris &

Raviv, 1978, p. 21; Shavell, 1979, p. 55; Arrow, 1985, p. 37; Petersen, 1993, p. 278). Theoutcome is usually somehow observable to both to the agent and the principal, and it

could also have many facets, such as quality and quantity. It can be, for example, thenumber of shoes produced by a factory worker, the volume of sales generated by a

department store salesperson, the success of a surgical procedure, and so forth.Environmental factors (a.k.a. random factors) such as economic conditions in the

market, competitors actions, and the weather are issues or conditions that arecustomarily beyond the agents and principals control. (Bergen et al., 1992, pp. 3-4;

Petersen, 1993, pp. 278 279.)

Most of the agency theoretical literature adopts two very important assumptionsconcerning the agency relationship. There must be (a) informational asymmetries and (b)

goal conflicts simultaneously present in the agency relationship (e.g. Moe, 1984, p. 754;Eisenhardt, 1989, p. 58; Barney & Hesterly, 1996, p. 125; Waterman & Meier, 1998, p.

177). Informational asymmetries can simply be considered as a claim that an agentpossesses more or better information about the details of individual tasks assigned to him,

his own actions, abilities and preferences (cf. Eggertsson, 1990, p. 41).

The level of informational asymmetries does not need to be stable; it can vary from onesituation to the next. The basic requirement is though, that the principals have to face

difficulties in acquiring the information possessed by the agent. The other assumption, ofgoal conflicts, can be understood as a situation where the principals and the agents

desires and interests concerning certain ends are in conflict with each other and that theywould therefore prefer different courses of action.

From the perspective of agency theory, goal conflicts arise because of the agents self-

interest and the tendency to maximize or pursue his individual utility. Conflicts betweenthe principals and the agents goals do not have to be permanent or constant, but there


28/74

must be a certain potential for them to occur (Milgrom & Roberts, 1992, p. 185).

2.8 The Roles of Agency Theory

Management

Agency theory has become a cottage industry that explores every permutation andcombination of agency experience in the corporate form. Because the work is largely

empirical, it by necessity relaxes some of the assumptions of classic agency theory ineconomics and business management even in the management literatures, specialized

academic and applied practitioner journals, the business press, even corporate proxystatements by the early 1990s, representing a new form and becoming the dominant

institutional logic of corporate governance (Zajac &Westphal 2004). Corporationsannounced the adoption of new policies, explicitly invoking agency theory buzzwords

about aligning incentives, discouraging self-interested behavior by managers, andreducing agency costs. Indeed, some adopted new policies that embraced an agency

rationale without bothering to implement them, simply jumping on the bandwagon of asocially constructed institutional logic that bestowed increased market value on symbolic

declarations alone (Zajac & Westphal 2004). Despite the fascinating case study in socialmovements (Davis & Thompson 1994), the diffusion of innovations, and the sociology of

knowledge that these developments offer, they also had a significant impact on theintellectual agenda of the academy, spawning a massive empirical literature in

management and organizational behavior. Agency theory has become a cottage industrythat explores every permutation and combination of agency experience in the corporate

form. Because the work is largely empirical, it by necessity relaxes some of theassumptions of classic agency theory in economics and business management. The most

popular stream of literature focuses on incentive alignment, particularly compensationpolicies. Empirical studies consider the types and correlates the literature also includes

studies of the process and costs of searching for agents, especially in light of the tensionsposed by adverse selection. Another major body of scholarship considers the agency

problems, agency costs, efficacy, and trade-offs of different control mechanisms as theyintersect and vary by

Length of principal-agent relationship; Organizational structure and form (e.g., headquarters and subsidiary,

outsourcing)

Characteristics of industries, organizations, and employees (e.g., technologies, Product demand, diversification, venture capitalist-entrepreneur relationships, Family firms, cultural distance between sites, employee education, Skill levels, amount of specialized knowledge, autonomy, etc.);


29/74

Programmability of the task, or how well the required behaviors can be Precisely defined (Eisenhardt 1989); and Organizational environments (e.g., turbulence).

Also coursing through this literature is a debate, sketched earlier, between those whoadopt the skeptical, even paranoid, assumptions of agency theory and the costly control

mechanisms it propounds and those who have a more hopeful view of human capacitiesfor other-regarding behavior and altruism and argue that agency costs can be mitigated by

organizational structures that foster reciprocity, cooperation, embeddedness, and trust(Donaldson 1990, Wright&Mukherji 1999).

Political Science

In exploring the delegation of power and authority in political and government

institutions and international organizations, political scientists take agency theory outsideof the economic marketplace and the constricting web of assumptions that shroud the

economic theory of agency. The political system can, of course, be understood as acomplex network of principal-agent relationships compose of citizens, nation states,

elected officials, lawmakers, members of the executive branch, administrative agencies,courts, international organizations, ambassadors, bureaucrats, soldiers, police officers,

supervisory officials, civil servants, patronage appointees, and even those who monitorother agency relationships inside political institutions and in the market. These actors

concurrently play principal and agent roles within and across political organizations.

A general theory of agency emerged in political science (Mitnick 1973) at the same timethat it did in economics (Ross 1973), apparently independently. As we have seen, the

latter took off spectacularly, becoming quickly institutionalized in an academic literature,specialty journals, and corporate ideologies and practices The former languished (Moe

1984), developing belatedly as rational choice theory made inroads into political science.As a result, agency theory in political science borrows heavily from the economics

paradigm rather than the more sociological conception offered by Mitnick (1973) or evenclassic works, such as Weber on bureaucracy (Kiser 1999).

The vague outlines of the agency paradigm in political science are the same as those in

the classic version: Principals delegate to agents the authority to carry out their politicalpreferences. However, the goals of principals and agents may conflict and, because of

asymmetries of information, principals cannot be sure that agents are carrying out theirwill. Political principals also face problems of adverse selection, moral hazard, and agent

opportunism. So principals contrive incentives to align agent interests with their own andundertake monitoring of agent behavior, activities that create agency costs.

The details are quite different, however, for many of the reasons considered earlier.

Political scientists assume multiple agents and principals; heterogeneous preferences orgoal conflict and competition among principals and among agents as well as between


30/74

them; problems of collective action; a more complicated palate of interests and thereforedifferent incentives mobilized to control them; varying sources of and mechanisms to

mitigate informational asymmetries; an active role for third parties (interest groups,regulated parties, etc.); and a dynamic playing field on which relationships unfold and are

transformed. Political scientists also consider a more diverse set of scenarios for

delegating power beyond those inherited from the economics paradigm. Principals madelegate to another to enhance the credibility of their commitments, for self-binding (toensure their long-term resolve in the face of immediate temptations), or to avoid blame

for unpopular policies.

These scenarios call for a very different agency contract. Instead of providing incentivesand sanctions to align the interests of agents with their own, principals seeking credibility

from their agents select agents operating at arms length, with very different policypreferences, and confer considerable discretion and autonomy to them. These agency

contracts grant independence while still seeking to insure accountability (Majone 2001).Early literature in political science on the iron law of oligarchy, the iron triangle (between

Congress, regulatory agencies, and regulated interests), regulatory capture, andbureaucratic drift all give voice to some of the intrinsic difficulties of principal control in

political institutions.

More recent work employing an agency theory perspective ranges from appellate reviewof lower court decisions to political corruption and presidential decisions to use force.

The largest literatures examine state policy implementation, the relationship betweenelective institutions and administrative agencies (especially legislators and bureaucrats),

and government regulation.

Principal-agent perspectives are also commonplace in examinations of internationalorganizations (e.g., central banks, international courts, the European Union) in the

literature on comparative politics and international relations. Political scientists devote farmore attention than economists to the details of how principals control agents. There is

some work on the selection and recruitment of agents, the role of patronage politicalappointments, and the impact of civil service requirements on adverse selection and more

on how principals specify their preferences.

A body of work considers statutory control (i.e., detailed legislation) and how lawmakerscraft legislation to restrict the discretion of those charged with its implementation,

specifying administrative structures and procedures to constrain the decision-makingprocess (McCubbins et al. 1989).

There are literatures on political oversight and monitoring, including ways in which

principals opt for reactive over proactive oversight, relying on third-party monitoring byaffected interest groups or the targets of their legislation to detect and report on

noncompliance (so-called fire alarms or decibel meters). There is more attention inpolitical science than in economics to the role of sanctions budget cuts, vetoing rules or

agency actions, reversing court decisions firing officials or voting them out of office,requiring agency reauthorization or threatening recontracting, etc. perhaps because, as


31/74

noted earlier, it is far less easy to align incentives without the financial inducements thatflow through economic organizations. The literature also considers the matter of agency

costs; when they are too high, principals may decide not to squander resources on them(Mitnick 1998, Banfield 1975). Because politicians may not directly feel the

consequences of self interested, opportunistic agents shirking or undermining their

interests (what political scientists call slack, slippage, or bureaucratic drift), the costs ofwhich are generally passed along to the public, monitoring activities may be more lax inpolitical arenas (Waterman & Meier 1998).

Law

Long before there was a theory of agency, there was a law of agency. Indeed, it was not

until the twenty-first century that the Restatement of the Law, Agency (American LawInstitute 2001) replaced master/servant with employer/employee. The law of agency

encompasses the legal consequences of consensual relationships in which one person (thprincipal) manifests assent that another person (the agent) shall, subject to the

principals right of control, have power to affect the principals legal relations through theagents acts and on the principals behalf (American Law Institute 2001, p. 1).

Agency doctrine defines the legal obligations that principals have with third parties for

action that agents took on their behalf. The principal, for example, may be bound tocontracts and transactions made by the agent and may be vicariously liable for some

instances of the agents misconduct (DeMott 1998, p. 1038). Because principals will beheld responsible for the actions of their agents, the law also attends to the sou

principal-agent final paper

Documents