The Uses and Abuses of Agency Theory in Business Ethics

The spectacular corporate scandals and bankruptcies of the past decade have served as a
powerful reminder of the risks that are involved in the ownership of enterprise. Unlike other patrons
of the firm, owners are residual claimants on its earnings. 1 As a result, they have no explicit contract
to protect their interests, but rely instead upon formal control of the decision-making apparatus of the
firm in order to ensure that their interests are properly respected by managers. In a standard business
corporation, it is the shareholders who stand in this relationship to the firm. Yet as the recent wave of
corporate scandals has demonstrated once again, it can be extraordinarily difficult for shareholders to
exercise effective control of management, or more generally, for the firm to achieve the appropriate
alignment of interests between managers and owners. After all, it is shareholders who were the ones
most hurt by the scandals at Enron, Tyco, Worldcom, Parmalat, Hollinger, and elsewhere. For every
employee at Enron who lost a job, shareholders lost at least US$4 million. 2 Furthermore, employees
escaped with their human capital largely intact. Creditors and suppliers continue to pick over the
bones of the corporation (which still exists, under Chapter 11 bankruptcy protection, and continues
to liquidate assets in order to pay off its debts).3 But as far as shareholders are concerned, their
investments have simply evaporated, beyond any realistic hope of retrieval. (In fact, one of the
reasons that Enrons collapse was particularly damaging to its employees was that so many of them
were also shareholders, through the company ESOP and their 401k plans.)

The sort of managerial attitude toward investors that proved so damaging in these scandals
was best illustrated by an internal memo written by Hollinger International CEO Conrad Black, who
described a self-dealing transaction conducted by Richard Perle, a member of the firms Board of
Directors, as containing a good deal of nest-feathering, yet complained only about the exclusion of
management from the benefits. They should treat us as insiders with our hands cupped as the
money flows down, and not as outsiders pouring in the money, he wrote. 4 It was the prevalence of
such attitudes toward shareholders (outsiders pouring in the money) that led an investigative
committee struck by the Board to later describe Perle as a faithless fiduciary, and Black as having
run a corporate kleptocracy5 At the time of writing Black is under criminal indictment; it is not
known whether he will be convicted for his role in several dubious transactions at Hollinger. Thus
there is still an open question as to whether he broke the law. It seems evident, however, that he
conducted himself unethically.

One of the central tasks of theoretical business ethics is to provide a conceptual framework
that allows us to articulate more precisely the intuitive sense we all have that nest-feathering and
similar forms of conduct are unethical, so that we can state more clearly the nature of the moral
obligations that have been violated. In approaching this task, the first place that business ethicists
1See Henry Hansmann, The Ownership of Enterprise (Cambridge, MA: Harvard University Press, 1992), p. 11.
2In the year 2000, Enron had 19,000 employees, and a peak market capitalization of $80 billion. Also, presumably not every
Enron employee had to search for work (hence the at least).
3It is estimated that creditors will get about 20 cents on the dollar. See Bethany McLean and Peter Elkind, The Smartest Guys
in the Room (New York: Penguin, 2003), p. 410.
4Perle Asserts Hollinger's Conrad Black Misled Him New York Times (Sept. 6, 2004).
5Panel Says Conrad Black Ran a 'Corporate Kleptocracy', New York Times (Aug. 31, 2004).

might reasonably be expected to look is to agency theory6 After all, the relationship between owners
and managers is a textbook example of a principal-agent relationship 7 Furthermore, deception and
misappropriation of funds by the agent represent perfect examples of the type of moral hazard
problems that are an endemic feature of principal-agent relations. Thus one might expect business
ethicists to embrace agency vocabulary as a way of stating with greater precision the exact nature of
the moral obligations that were violated at Enron and elsewhere. 8 One might also expect business
ethicists to insist that greater

attention be paid to agency relations, and to the potential moral hazard problems that they harbor,
as a way of avoiding such scandals in the future. Indeed, many have done so.

However, the reaction to the scandals among business ethicists has been far more mixed than
one might expect. Part of the reason is that many business ethicists have spent considerable time and
energy downplaying the importance of shareholders in the organizational structure of the firm, and
trying to show that managers have important moral obligations to other stakeholder groups. 9 Many
deny that managers should be regarded as agents of the shareholders in any significant sense of the
term. Thus they do not regard the recent spate of corporate scandals as grounds for renewed attention
to the agency risks that exist in the manager-shareholder relation. On the contrary, some have gone
so far as to blame agency theory - and the teaching of agency theory in business schools - for
creating the corporate culture that led directly to the scandals. Rakesh Khurana, Nitin Nohria, and
Daniel Penrice of the Harvard Business School have suggested that the doctrine of shareholder
primacy combined with agency theory led directly to many of the worst profit-maximizing abuses
unmasked in the recent wave of corporate scandals.10 Along similar lines, Brian Kulik has argued
that agency reasoning on the part of Enron executives led to the creation of an agency culture
and an organizational structure within the firm that encouraged corrupt behavior.11

So which is it? Is agency theory a part of the problem, or a part of the solution? In order to
get clear on this question, it is important first to get clear on the sort of theoretical commitments that
are
6For a complete technical overview of this theory, see Jean-Jacques Laffont and David Martimort, The Theory of Incentives:
The Principal-Agent Model (Princeton: Princeton University Press, 2002).
7See Paul Milgrom and John Roberts, Economics, Organization and Management (New Jersey: Prentice Hall, 1992), p. 170;
Donald E. Campbell, Incentives (Cambridge: Cambridge University Press, 1995), pp. 79-86.
8Allen Buchanan has provided what is perhaps the most sophisticated development of this approach. See Toward a Theory of
the Ethics of Bureaucratic Organizations, Business Ethics Quarterly, 6 (1996): 419-440.
9See, e.g., Marjorie Kelly, Why all the Fuss about Stockholders? reprinted in her The Divine Right of Capital (San
Francisco: Berrett-Koehler, 2001); Max Clarksons introduction to The Corporation and its Stakeholders (Toronto: University
of Toronto Press, 1998); Margaret M. Blair, Corporate ownership: A Misleading Word Muddies the Corporate Governance
Debate, The Brookings Review (Winter 1995): 16-19.
10Rakesh Khurana, Nitin Nohria and Daniel Penrice, Management as a Profession in Jay W. Lorsch, Leslie Berlowizt, and
Andy Zelleke, eds., Restoring Trust in American Business (Cambridge, MA: MIT Press, 2005). Some theorists, such as Dennis
P. Quinn and Thomas M. Jones, An Agent Morality View of Business Policy, Academy of Management Review, 20 (1995):
22-42, simply equates agency theory with the doctrine of shareholder primacy, which leads then to the suggestion that anyone
committed to stakeholder theory must reject agency theory.
11Brian W. Kulik, Agency Theory, Reasoning and Culture at Enron: In Search of a Solution, Journal of Business Ethics, 59
(2005): 347-36.

essential to agency theory (in order to distinguish between agency theory itself and certain incorrect

interpretations that have become widely promulgated). It is also important to be more specific about
the ways that agency theory can be used to analyze relations within the firm, in order to determine
whether it is the use or the abuse of agency theory that has become a source of mischief. Finally, it is
important to be more specific about the circumstances in which moral obligations can arise out of
agency relations. Only then is it possible to develop a more balanced appreciation of the contribution
that agency theory can make to the study of business ethics.

1; What is agency theory?

Agency theory, in the sense that the term is used here, is an approach that involves the
application of game theory to the analysis of a particular class of interactions, viz. situations in
which one individual (the agent) acts on behalf of another (the principal) and is supposed to advance
the principals goals.12 This is already a potential source of confusion, since the term agent is used
differently here than in certain other contexts, such as corporate law, where the law of agency
assigns a much narrower meaning to the term.13 But these disputes over the use of the term agent
are not where the real controversy arises. It is the use of game theory that makes agency theory
controversial. This is because game theory comes freighted with a number of substantive theoretical
assumptions, including most prominently, a commitment to an instrumental (or economic) model
of rational action. Thus individuals are represented as expected utility-maximizers (who, when faced
with a problem of interdependent choice, select actions that represent an individually best response
to the anticipated actions of the other individuals). This immediately raises the dander of many
ethicists, since economic models of rationality are famous for either classifying all moral action as
irrational, or else rationalizing it through the discovery and ascription of some underlying nonmoral incentive.

Thus ethicists often complain that agency theorists, by adopting an economic model of
action, thereby assume that rational individuals are self-interested, or that they act only from egoistic
and not altruistic motives. This is, from their point of view, equivalent to endorsing moral
skepticism, and is therefore not a helpful point of departure for the development of a system of
applied ethics. Of course, the standard response to this criticism is to say that the economic model of
rationality implies no such thing. Utility is defined with respect the preferences of individuals, and
preferences reflect whatever desires individuals happen to have, egoistic or altruistic. 14 David
Gauthier made the point most succinctly, when he observed that, in the economic model of
12Milgrom and Roberts, Economics, Organization and Management, p. 170.
13 See Robert C. Clark, Agency Costs Versus Fiduciary Duties, in John W. Pratt and Richard J. Zeckhauser, Principals
and Agents (Boston: Harvard Business School Press, 1985), p. 56. In part because of this potential for confusion, some
game theorists have begun to speak more generally about the theory of incentives. See, e.g. Laffont and Martimont,
The Theory of Incentives, also Campbell, Incentives.
14For an example of a typical - and typically acrimonious - exchange between a business ethicist and a game theorist on this
point, see Robert C. Solomon, Game Theory as a Model for Business and Business Ethics, Business Ethics Quarterly, 9
(1999): 11-30, and Ken Binmore Game Theory and Business Ethics, Business Ethics Quarterly, 9 (1999): 31-36.

rationality, it is not the interests in the self, that take oneself as object, but interests of the self, held
by oneself as subject, that provide the basis for rational choice and action. 15 Thus what creates the
need for incentives in principal-agent relations, strictly speaking, is not the fact that the principal and
the agent have egoistic preferences, but merely the fact that they have different preferences.
Principal-agent theory is about how individuals manage situations involving goal incongruity
between two or more persons.16 It does not matter whether they are selfish or not; what matters is
that each acts in pursuit of his or her own goals, and that the goals of the other show up only insofar
they affect that agents goals, or ability to satisfy these goals.

On these grounds, many business ethicists have concluded that agency theory is perfectly
anodyne. Allan Buchanan articulates this view well when he writes:

If, in applying principal/agent theory, it were necessary to assume that motivation is

exclusively

or primarily self-interested, this would greatly reduce if not vitiate the enterprise. However,
we need not do so. Instead, we can proceed on the assumption that the conflicts of interest
that give rise to agency-risks may result from a variety of motivations, on the part of agents
and principals. All that is necessary is that there be conflicts of interest.17

Of course, in fairness to those business ethicists who have complained about the self-interest
assumption, it should be noted that one can search the economic theory of the firm literature for a
very long time before finding an actual example of an agency analysis that ascribes altruistic motives
to any of the parties involved. Even if the theoretical framework does not force them to do so,
agency theorists often do make unflattering empirical assumptions about individual preferences, by
stipulating in their models that, for example, work effort has negative utility, money rewards have
positive utility, and that individuals have no other relevant motives. 18 Strictly speaking, however,
such assumptions are not essential to the economic model of rationality, and so theorists like
Buchanan are quite correct to point out that agency theory per se entails no commitment to such
claims.

It would be premature, however, to conclude on this basis that the economic conception of
rationality is neutral from the standpoint of ethics. There are a number of other substantive
theoretical commitments associated with the instrumental model, which are hostile from the
perspective of the ethicist, and which cannot be purged from the model so easily.
15David Gauthier, Morals by Agreement (Oxford: Clarendon, 1986), p. 7. See also Daniel M. Hausman and Michael S.
McPherson, Economic Analysis and Moral Philosophy (Cambridge: Cambridge University Press, 1996), pp. 52-53.
16J. Gregory Dees, Principals, Agents and Ethics, in Norman E. Bowie and R. Edward Freeman, eds. Ethics and Agency
Theory (New York: Oxford University Press, 1992), pp. 37-38.
17Buchanan, Toward a Theory of the Ethics of Bureaucratic Organizations, p. 421.
18J. Gregory Dees, Principals, Agents and Ethics, p. 29.

The first of the two outstanding problems stems directly from the tendency among game
theorists to black box all questions of motivation. While this theoretical strategy does allow them
to sidestep disputes over altruism and egoism, it also leaves them without a developed theory of
preference-formation, and thus without any ability to model the way that preference changes arise
out of social interactions.19 Preferences are taken as given, and are also taken to be independent of
strategies. Thus players in a standard game-theoretic model cannot change each others preferences
through their actions. This is closely related to the fact that in standard game-theoretic models
players are explicitly precluded from communicating with one another (using any sort of
independent semantic resources, such as language; they are still able to draw inferences from
observing each others actions, and so are able to communicate in this sense). 20 Furthermore,
insofar as they are able to communicate with one another, standard game-theoretic solution
concepts, like Nash equilibrium, do not apply.21 This non-trivial restriction on game-theoretical
models is often conveniently forgotten by those who are eager to apply them to the analysis of
empirical interactions.

In any case, the fact that there is no generally accepted or robust theory of endogenous
preference-change in games means that agency theorists have devoted almost all of their time and
attention to studying the way that external incentives can be used to bring about greater alignment of
goals in cases of incongruity. This often turns into a classic case of economists searching whether the
light is best. For instance, in their widely-used management textbook on organizational theory, game
theorists Paul Milgrom and John Roberts dedicate an entire chapter to the subject of moral hazard
and agency relations within the firm. They canvas an exhaustive range of strategies for controlling
employee shirking, including monitoring, incentive contracts, performance pay, ownership stakes,
employee bonding, and promotional systems. At the same time, they fail to mention such absolutely
elementary factors as whether or not employees enjoy their jobs, and whether they love or hate the
firm that they work for.22 Similarly, in their chapter on human resources policy, Milgrom and Roberts
have a lengthy discussion of employee retention strategies, which does not once mention the fact that
employees sometimes feel a sense of loyalty toward the firm (and that managers have it within their
power to cultivate such loyalties). On occasion, this occlusion of motivational issues borders on the
comical, as when they develop a case study of human resources policies in Japan that manages to
avoid mentioning the issue of employee loyalty altogether. The control structure of Japanese firms,
which gives considerable power to the employees as a group is explained, not as a way of
promoting loyalty and building esprit de corps, but rather as a way of enabling employees to
protect their valuable employment rights in the face of labor-market rigidity.23 It is greater fear of
losing their jobs, we are led to believe, that makes Japanese workers more willing than Americans to
accept sacrifices on behalf of their employer.24

Once again though, this emphasis on external incentives is not a necessary consequence of
19Jack Knight, Institutions and Social Conflict (Cambridge: Cambridge University Press), p. 18.
20See John Nash, Non-cooperative Games, Annals of Mathematics, 54 (1951): 289-295.
21Joseph Heath, Communicative Action and Rational Choice (Cambridge, MA: MIT Press, 2001), pp. 73-78.
22Milgrom and Roberts, Economics, Organization and Management, pp. 179-192. In this respect, their discussion falls
significantly below the level of sophistication exhibited in older classics such as James G. Marsh and Herbert Simon,
Organizations (New York: John Wiley & Sons, 1958), pp. 65-81.
23Milgrom and Roberts, Economics, Organization and Management, p. 350.
24Compare this view to the discussion in Francis Fukuyama, Trust (London: Penguin, 1995), pp. 185-193, 255-266.

the commitment to the economic conception of rational action. There is nothing intrinsic to agency
theory that prevents people from taking an interest in the way that internal incentives - e.g.
preference change - can be used to overcome agency problems, it is just that game theorists have no
idea how to model such processes, and so have largely chosen to ignore them (in very much the
same way that, prior to the advent of game theory, economists had no way to model information
states, and so chose to ignore the impact of asymmetric information on market exchanges). Thus the
emphasis on external incentives is simply a case of methodologically induced bias, which could be
corrected through the development of more sophisticated modeling techniques - or even just a frank
acknowledgment of the need for qualitative analysis in this domain. So again, there is no reason in
principle for the ethicist to object to the use of agency theory.

The second outstanding problem, however, has no quick fix. It involves the commitment, on
the part of the agency theorist, to the view that individuals will behave opportunistically whenever
given

the chance to do so.25 For example, it is routinely assumed that regardless of what people say they
are going to do, they will always update their plans as the situation unfolds, and renege on any prior
commitments whenever it is in their interest to do so. Thus a farmer may hire workers who promise
to harvest his crop, but find himself facing a strike threat at a critical time during the season, when it
is too late to bring in replacement workers.26 An insurance company may agree to indemnify any
policyholder who suffers a particular sort of loss, but then drag its feet when the time comes to pay
the claim (e.g. by proposing unusual legal interpretations of certain exclusion clauses). Employees
may agree to give some particular job their full attention, but then shirk in various ways in situations
where their effort level is unobservable, and so on.

Along with this characterization of opportunistic behavior comes the assumption that
individuals are unable to credibly commit themselves to refraining from opportunistic behavior,
unless they are able to create some external incentive structure that changes their own future
incentives (such as posting a bond to guarantee performance). Promises to perform are basically
cheap talk, and the rational principal will disregard them when it comes to managing agency
relations.

Ethicists are unlikely to regard this as a satisfactory framework for analysis, since it suggests
that rationality requires individuals to exhibit a variety of vices, including fickleness (in
Machiavellis sense of the term), dissimulation, treachery and guile. It also follows very closely upon
this that rational agents will treat each other with distrust and suspicion. Thus agency theory seems
to take some of the worst assumptions about human nature and build them into its central definition
of rationality. Furthermore, in this case the standard evasive response is not available to the agency
theorist. Unlike the egoism postulate, which is in fact peripheral to the instrumental conception of
25For definition and discussion of this term, see Oliver E. Williamson, The Economic Institutions of Capitalism (New York:
The Free Press, 1985), pp. 47-49.
26Milgrom and Roberts, Economics, Organization and Management, p. 128.

rationality, the assumption of opportunistic behavior is absolutely central to the model. The fact that
agents are unable to make commitments is one of the defining postulates of non-cooperative game
theory (and again, all of the standard solution concepts do not apply in cases where that assumption
is relaxed).27 What we typically refer to as opportunistic behavior is a direct consequence of agents
acting in accordance with the general game-theoretic principle known as sequential rationality. This
is simply the view that, in a multi-stage game, a rational strategy must not only be utilitymaximizing at the point at which it is chosen, but each of its component actions must also be utilitymaximizing at the point at which it is to be performed. The sequential rationality postulate is what
licenses, among other things, the use of backward induction as a method for solving multi-stage or
repeated games.28 It is so deeply entrenched that, in most cases, game theorists dont even bother to
mention it. Eric Rasmusen, for example, in his widely-used textbook on game theory, discusses the
principle only once, in order to explain why he will not be mentioning it again:

The term sequential rationality is used to denote the idea that a player should maximize his
payoffs at each point in the game, re-optimizing his decisions at each point and taking into
account the fact that he will re-optimize in the future. This is a blend of the economic idea of
ignoring sunk costs and rational expectations. Sequential rationality is so standard a criterion
for equilibrium now that often I will speak of equilibrium without the qualifier when I
wish to refer to an equilibrium that satisfies sequential rationality...29

Opportunism, from this perspective, is just a somewhat moralizing way of describing the
phenomenon of re-optimization, and as such, is not easy to get rid of as a game-theoretic
assumption. On the contrary, it comes very close to capturing the essence of the strategic conception
of rationality.

Central to this conception is the consequentialism postulate, which states simply that the value of an
action is a functions of its anticipated consequences, and nothing else (the commitment to reoptimization follows almost immediately from this consequentialism). Yet consequentialism
precludes the possibility that a rational agent might incorporate deontic constraints - principles
associated directly with actions, independent of their consequences - into his or her deliberations. 30
Since genuine loyalty, commitment, conformity to social norms and respect for moral rules are all
forms of deontic constraint, this is a very significant restriction. Thus when a critic like Eric Noreen
claims that at the heart of agency theory, as expounded in accounting, finance and economics, is the
assumption that people act unreservedly in their own narrowly defined self-interest with, if
necessary, guile and deceit,31 he is only half wrong. While it is incorrect to say that self-interest,
narrowly defined, is at the heart of agency theory, it is correct to associate agency theory with the
view that people act unreservedly, using guile and deceit - not even when necessary, but whenever it
is advantageous for them to do so. Thus the image of employees loafing around whenever the boss
27Heath, Communicative Action and Rational Choice, pp. 86-92.
28Drew Fudenberg and Jean Tirole, Game Theory (Cambridge, MA: MIT Press, 1991), pp. 72-74
29Eric Rasmusen, Games and Information, 3rd edn. (Oxford: Blackwell, 1989), p. 95.
30Or what Robert Nozick refers to as side constraints, Anarchy, State and Utopia (New York: Basic Books, 1974), pp.
28-32.
31Eric Noreen, The Economics of Ethics: A New Perspective on Agency Theory, Accounting, Organizations and
Society, 13 (1988): 359-369.

isnt looking, faking disabilities, calling in sick during hunting or fishing season, exaggerating the
difficulty of their assignments in order to make their performance appear more impressive, and so
on, is a non-accidental consequence of the agency perspective.32

Thus business ethicists do have some legitimate concerns about the agency theory
framework, insofar as it incorporates a controversial conception of rationality, one that presupposes
the correctness of a certain form of moral skepticism. Yet even then, it is unclear that these concerns
need ripen into full-blown complaints. After all, most agency theorists are not in the business of
doing normative theory. In other words, they are not telling people how they should behave (and thus
are not directly recommending opportunism, guile and deceit as laudable forms of behavior). Their
goal typically has been to develop a positive theory of the firm, to offer merely empirical
explanations of why organizations take on particular forms, structured by particular sets of
incentives. If, in doing so, they make certain unflattering assumptions about human behavior, why
should that be any cause for alarm among ethicists? And how could the mere teaching of such
methodological tools in business schools be blamed for the recent spate of corporate scandals?

2;

Whats the problem with agency theory?

Here is the question, simply put: if agency theory is merely a tool used to develop a positive
theory of the firm, how much mischief could it really cause in a corporate environment? The answer
is: more than one might think.

The first step to understanding this answer lies in an appreciation of the fact that, as a
positive theory of the firm, agency theory generates predictions that are wildly at variance with what
one can actually observe in the behavior of individuals and in the structure of organizations. In other
words, it generates a positive theory that, insofar as it is falsifiable, is demonstrably false. Of course,
many of the potential problems identified by agency theory are no doubt genuine - this is why the
theory resonates with so many people. There is, for example, a notable tendency toward moral
hazard. Similarly, individuals have a tendency to act non-cooperatively in collective action problems.
Usually, however, these show up only as tendencies, even when game-theoretic analysis predicts
universal defection. In particular, while moral hazard in the firm can be a serious problem,
empirically it is much less of a problem than any straightforward application of game-theoretic
analysis to principal-agent relations would lead one to predict.

The empirical limitations of game-theoretic models have, of course, been exhaustively

studied and documented by experimental game theorists. It is well-known, for instance, that large
numbers of individuals cooperate in one-shot prisoners dilemmas, knowing full well that there is no
possibility of reciprocation. This fairly large-scale deviation from the equilibrium strategy is not a
blip or an artifact of some particular experimental procedure - cooperation remains stable under a
32Most of these examples are drawn from Milgrom and Roberts, Economics, Organization and Management, p. 170.

wide variety of conditions: across a wide range of different cultures, among subjects playing for the
first time and among those with previous experience, in groups ranging from 4 to 80 members, and
with a variety of different monetary rewards.33

Apart from the prisoners dilemma, the other game that has been widely studied in
experimental settings is the ultimatum game. Here, one player is given a fixed sum of money and
told to propose some division of the money between himself and one other person. The second
player can then either accept this proposal, in which case the money is divided up as per the offer, or
reject the proposal, in which case both players receive nothing. Of course, the second player never
has any positive incentive to reject any offer, since no proposed division is worse than receiving
nothing. Thus rejecting the offer is a punitive action - and the threat to do so is precisely the sort of
commitment that sequential rationality rules out. As a result, standard game theory suggests that the
proposer should select a division that gives the second player as little as possible (in a sense,
behaving opportunistically), and that this proposal should always be accepted. In reality, not only do
players tend to offer much more than the instrumental analysis predicts, but proposals also tend to be
rejected if they fall too low. In industrialized societies, mean offers tend to be around 44 per cent,
while offers below 20 per cent are rejected 40 to 60 per cent of the time. Experimental evidence
from non-industrialized societies reflects greater variability - including examples of mean offer rates
above 50 per cent, combined with frequent rejection of such offers. Yet in spite of these variations,
no experiment has ever come close to conforming to the expectations of standard game theory.34

Given these experimental findings, it would not be surprising to find that agency theory
consistently overstates the agency costs that may arise within organizations, simply because real
human beings often behave cooperatively, exhibit loyalty, and refrain from acting opportunistically,
even in the absence of external incentives. This fact is of course well-understood by sophisticated
management theorists, even those deeply wedded to the agency perspective. The general upshot of a
lot of agency analysis of the firm is that many organizations, especially those that exhibit what
Oliver Williamson calls information impactedness, simply would not function if the only tools that
managers had at their disposal were external punishments and rewards. 35 Bengt Holstrom showed
very early on how imperfect observability could make it impossible to devise efficient incentive
schemes for individuals working in teams.36 George Baker and others drew attention to the fact that,
when effort or output was not fully observable, a system of sharp incentives focused upon one aspect
of the task could produce results that were much worse than a system of dull incentives applied to
the task as a whole.37 Much of the agency literature wound up sounding a very skeptical note on the
subject of performance pay, and provided unexpected support for the old-fashioned practice of
33Robyn M. Dawes and Richard H. Thaler, Anomalies: Cooperation, The Journal of Economic Perspectives, 2 (1988): 187197. See also Friedrich Schneider and Werner W. Pommerehne, Free Riding and Collective Action: An Experiment in Public
Microeconomics, Quarterly Journal of Economics, 96 (1981): 689-704, Oliver Kim and Mark Walker, The Free Rider
Problem: Experimental Evidence, Public Choice, 43 (1984): 3-24, Mark Isaac, Kenneth F. McCue, and Charles R. Plott,
Public Goods Provision in an Experimental Environment, Journal of Public Economics, 26 (1885): 51-74.
34Joseph Henrich, Robert Boyd, Samuel Bowles, Colin Camerer, Ernst Fehr, Herbert Gintis, and Richard McElreath,
Cooperation, Reciprocity and Punishment in Fifteen Small-scale Societies, American Economic Review, 91 (2001): 73-78.
The offers above 50 per cent are typical in societies with gift economies, or legacies thereof.
35Oliver Williamson, Markets and Hierarchies: Some Elementary Considerations, American Economic Review, 63 (1973):
316-325, at 318.
36Bengt Holstrom, Moral Hazard in Teams, Bell Journal of Economics, 13, (1982): 324-340.
37George P. Baker, Incentive Contracts and Performance Incentives, Journal of Political Economy, 100 (1992): 598614.

paying employees a flat salary.38 Results such as these suggested that, insofar as real-world
corporations do actually succeed in extracting reasonable levels of cooperative effort from their
employees, there must be more than just external incentives at work.

Given these results, one might wonder where the harm could be in business schools teaching
agency theory, or in managers using it as an analytic tool. And perhaps there would be no problem,
except for the fact that the limitations of the theory are often overlooked or understated. This can
lead to mischief in several different ways:

1. Imputed incentives. People who are overly impressed by economic methodology often subscribe
to the instrumental conception of rationality in a form that makes the model essentially unfalsifiable.
As a result, when particular agency problems do not show up where agency theory predicts that they
should, rather than concluding that there must be some relevant sort of internal constraint at work,
these theorists assume that the external incentives must be there, but that they simply have not been
discovered yet. Economists have in fact invested extraordinary ingenuity and effort in the task of
devising baroque external incentive schemes as a way of explaining phenomena that in fact admit of
far more straightforward internal explanations. To take just one example, there are two prominent
interpretations of the so-called efficiency wage phenomenon. Henry Ford set the relevant
precedent, by voluntarily increasing the pay of his workers to $5 a day at a time when average wages
in the automobile industry were less than half that. He was rewarded with a significant increase in
worker productivity (so much so that he later described it as one of the finest cost-cutting moves we
ever made.39) The commonsense explanation would be to suppose that Ford tapped into an
underlying norm of reciprocity.40 According to this perspective, the notion of a fair days work for a
fair days pay plays a powerful role in determining employee effort levels (or the old Soviet variant
we pretend to work, and they pretend to pay us).41 So when the boss agrees to pay you a rate that
it is, by common admission, far in excess of what he is obliged to pay, he has in essence done you a
favor. And since

one good turn deserves another, you then owe it to him to put more effort into your work (or at
very least, to refrain from shirking). One might also expect this obligation to be enforced
informally in the relations between workers on the shop floor, thus removing an important barrier
to observability and leading to a dramatic reduction in moral hazard problems.

It should also be noted that, apart from its commonsense appeal, there is significant
empirical evidence to support this norm of reciprocity explanation of efficiency wages. 42
38For an overview, see Robert Gibbons, Incentives in Organizations, Journal of Economic Perspectives, 12 (1998): 115132.
39Henry Ford, My Life and Work (Garden City, NY: Doubleday, 1922), p. 147.
40For example, see E. Fehr, S. Gachter and G. Kirchsteiger Reciprocal fairness and Noncompensating Wage Differentials,
Journal of Institutional and Theoretical Economics, 152 (1996): 608-640. George Akerloff, Labour Contracts and Partial Gift
Exchange, Quarterly Journal of Economics, 84 (1982): 488-500.
41Hausman and McPherson, Economic Analysis and Moral Philosophy, pp. 55-56.
42For overview, see discussion in Uri Gneezy, Do High Wages Lead to High Profits? An Experimental Study of Reciprocity
Using Real Effort (forthcoming).

1
0

Nevertheless, many economists have felt the need to resist this explanation. The more popular
suggestion has been that, by paying workers an above-market wage rate, Ford essentially created an
economic rent associated with employment at his firm. This made workers more averse to losing
their jobs, by making it unlikely that they would find work at comparable wages elsewhere. This,
combined with the queues of workers that began to assemble outside Fords factory looking for
work, created enough fear of dismissal to motivate the existing workers to shirk less. 43 According to
this view, the efficiency effects of the wage increase can be explained entirely through reference to
traditional monetary incentives, and without appeal to any obscure internal motivational factors,
such as a sense of fairness or a commitment to reciprocity. (Of course, few people would doubt that
the external explanation represents a part of the story, perhaps even an important part. The
question is whether it represents the entire story.)

John Boatright has argued that this methodologically-induced bias toward explanations in
terms of external incentives can have a psychological framing effect that, when translated into
practical managerial decision-making, might result in mistaken solutions to problems or even
incorrect assessments of the problems to be solved. 44 For example, the agency perspective is apt to
lead to a

distrust of agents and a reliance on mechanisms of control. Such an approach is warranted in certain

situations, but when applied in a business setting it may result in an overinvestment in monitoring
and other contractual solutions and a corresponding underinvestment in building trust in an
organization, and in fostering traits like loyalty and professionalism.45

The other potential source of mischief is caused by the assumption that, whenever a
particular sort of agency cost fails to arise, there must always be an explanation in terms of external
incentives. This can encourage individuals in such agent positions to act in a purely instrumental
(or self- interested) fashion, by leading them to assume that there must already be a system of
checks and balances in place to mitigate the negative impact of any opportunistic actions that they
take, even if they cannot see it. If they believed, on the other hand, that the situation called for moral
restraint on their part, as the only way of avoiding an agency cost or a collective action problem,
then they might be less willing to act opportunistically or non-cooperatively. They would certainly
be deprived of one powerful rationalization for unethical conduct.

43Timothy C.G. Fraser and Robert G. Waschik, Managerial Economics: A Game-Theoretic Perspective (London: Routledge,
2002), p. 291.
44John Boatright, Ethics in Finance (Oxford: Blackwell, 1999), p. 48. See also J. Gregory Dees, Principals, Agents and
Ethics, in Norman E. Bowie and R. Edward Freeman, Ethics and Agency Theory (Oxford: Oxford University Press, 1992), p.
35.
45Boatright, Ethics in Finance, p. 49. For further discussion, see papers collected in Bruno S. Frey and Margit Osterloh, eds,
Successful Management by Motivation - Balancing Intrinsic and Extrinsic Incentives (Berlin: Springer Verlag, 2002).

1
1

For example, many agency theorists downplay the significance of the fiduciary
relationship that exists between managers and shareholders.46 A fiduciary relationship implies both a
duty of care and a duty of loyalty, both concepts that are unintelligible as such within a gametheoretic framework. Thus there is a tendency among agency theorists to resist taking these
obligations at face value, and to regard them instead as just legal shorthand for a certain set of
implicit contracts, ones that are ultimately structured by external incentives. 47 But such an analysis
can easily lead those who are in a fiduciary role to take these obligations less seriously, and to act in
a more self-interested fashion, on the grounds that these implicit contracts already anticipate such
forms of behavior. When combined with the so-called efficient markets hypothesis, which
dramatically underplays the information asymmetry between shareholders and managers, the result
can be a very straightforward rationalization of unethical conduct.

Robert Clark describes the basis of this rationalization as a facile optimism about the
optimality of existing institutions.48 For example, it is common among those who share the
implicit contracts perspective to regard management nest-feathering, not as a breach of fiduciary
duty, but merely as implicit compensation. Managerial misrepresentation of company accounts (i.e.
loose accounting standards) is sometimes defended, and opportunistic behavior is excused, on the
grounds that it must have already been implicitly accounted for. Consider the following argument,
due to Lawrence Revsine:

It is reasonable to presume that those who negotiate managers employment contracts

anticipate such opportunistic behavior and reduce the compensation package accordingly.
Notice that the demand for loose standards is further increased insofar as managers bear
some or all of the agency costs. Since they have already be charged for the anticipated
opportunistic actions, they must now engage in them in order to achieve the benefits they
paid for. Since loose standards facilitate opportunistic actions, the demand for such
standards increases.49

Here the facile optimism about efficient contracting is presented, not just in a way that
excuses breaches of fiduciary duty, but in a way that actually puts pressure on managers to violate
these duties. After all, if a manager has already been charged for a padding an expense account, in
the form of reduced compensation, then he or she would be a fool to refrain from padding it. More
generally, any manager who does not take advantage of any and all opportunities for opportunism
is essentially being suckered.

46For example, see Michael C. Jenson and William J. Meckling, Theory of the Firm: Managerial Behavior, Agency Costs and
Ownership Structure, Journal of Financial Economics, 3 (1976): 305. For critical discussion on this point, see Robert C.
Clark, Agency Costs versus Fiduciary Duties, p. 59-60.
47Frank H. Easterbrook and Daniel R. Fischel write that, the fiduciary principle is an alternative to elaborate promises and
extra monitoring. It replaces prior supervision with deterrence, much as criminal law uses penalities for bank robbery rather
than pat-down searches of everyone entering banks, The Economic Structure of Corporate Law (Cambridge, MA: Harvard
University Press, 1991), p. 92. Notice how this redescription downplays the element of internal constraint in a fiduciary
relation, arguing instead that it involves simply a reshuffling of external incentives.
48Clark, Agency Costs or Fiduciary Duties, p. 65.
49Lawrence Revsine, The Selective Financial Misrepresentation Hypothesis, Accounting Horizons, 5 (1991): 16-27 at 18.

1
2

From an ethical perspective, the impact that such reasoning can have should not be
underestimated. The idea that ill-gotten gains are merely implicit compensation is one of the most
important techniques of neutralization used by white-collar criminals to rationalize - and hence to
grant themselves permission to engage in - illegal conduct. 50 Thus one can see in Revsines argument
a clear example of how a false understanding of agency theory and its implications can serve as a
powerful impetus toward both immoral and illegal behavior. Of course, there is a sense in which
agency theory itself is not to be blamed. Nevertheless, this false understanding is extremely
widespread, and so the potential for mischief that it creates merits emphasis.

2. Crowding out of moral incentives. As we have seen, the methodological biases of agency theory
generate an overemphasis on external incentives as a way of addressing agency risks, along with a
comparative neglect of internal incentives. Thus an enormous amount of time and energy has been
frittered away designing increasingly clever incentive schemes, to the neglect of more obvious
strategies for securing employee loyalty and dedication. Yet while this may be a waste of time, one
might also be inclined to think that it also can do no harm. Even if an organization depends heavily
upon voluntary deontic constraint on the part of its employees in order to avoid certain potential
agency problems, surely it cant hurt to layer on some additional external incentives, in order to
create a greater alignment of interests?

Of course, the agency literature itself is full of cautionary examples of how incentive
schemes can distort incentives, and thus of how poorly designed incentives schemes can exacerbate
agency

problems. Yet there is a more general problem that has been almost entirely ignored, namely, that
even a well-designed system of external incentives has the potential to undermine moral motivation,
and thus to create agent costs where previously none existed.51 Bruno Frey and Felix Oberholzer-Gee
refer to this phenomenon as the crowding out of moral incentives. Their research highlights some
of the ways in which pecuniary incentives can have the effect of undermining moral incentives. In
one study, they examined the willingness of citizens to accept NIMBY (Not In My Backyard)
projects, such as nuclear waste disposal sites, in Switzerland. 52 Nuclear power plants produce
benefits that are enjoyed quite widely, but impose highly localized costs (such as the dangers
associated with waste storage and disposal). This gives local communities an incentive to free ride
- to use the electrical power, but then refuse to accept either generation or disposal facilities in their
region. Frey and Oberholzer-Gee found that in one Swiss village that had been identified by experts
as the best disposal site, a slender majority of citizens (50.8 per cent) were willing to accept the
creation of such a facility in their community (and thus to act cooperatively). Yet surprisingly,
when officials decided to sweeten the offer by adding fairly large sums of money as compensation
(US$2,175-$6,525), support for the project plummeted to 24.6 per cent.
50James William Coleman, Toward an Integrated Theory of White-Collar Crime, American Journal of Sociology, 93 (1987):
406-439, at 414.
51See Ernst Fehr and Simon Gachter. 2002. Do Incentive Contracts Crowd Out Voluntary Cooperation? University of
Zurich Working Paper 34. [Link]
52Bruno S. Frey and Felix Oberholzer-Gee, The Cost of Price Incentives: An Empirical Analysis of Motivation CrowdingOut, American Economic Review, 87 (1997): 746-755.

1
3

It is not difficult to imagine what went on. Based upon simple cost-benefit calculation, it is
very unlikely that any community would find it in their interest to accept a nuclear waste disposal
facility. The value of the power that they (along with everyone else) receive is simply not worth it,
especially when there is some chance that concerted resistance to the project will result in its being
located in some other community (i.e. that NIMBY free-riding is a feasible option). Thus monetary
compensation is not likely to tip the balance for many people. The only way to get citizens to accept
such a facility is through a moral appeal, which might lead them to overlook their self-interest in
favor of the greater good. When considering the project from the moral point of view, citizens
simply do not engage in the relevant cost-benefit calculations. They approach the question from the
standpoint of what John Rawls calls the reasonable, rather than the rational. 53 Furthermore, their
consent may be based upon the fact that they do not enter into these calculations. (This interpretation
is suggested by the fact that willingness to accept the nuclear-waste facility was highly correlated
with abstract support for nuclear power as a means of electricity generation. Unless individuals are
basing their decisions upon a principle, this correlation is difficult to explain, since proponents of
nuclear power have the same freerider incentive as opponents. Furthermore, the impact that the
money offer had upon citizens perceptions of the risk associated with the project was negligible.)
Offering people external incentives has the effect of changing their perspective, so that they no
longer consider the question from the moral point of view, but rather examine it from the standpoint
of their self-interest. If the external incentives are inadequate, then the incentive scheme may easily
have the effect of undermining cooperation, thereby creating real collective action problems where
previously there were only potential ones.

What this shows is that internal and external incentives are not necessarily complementary or
cumulative, even when in theory they are correctly aligned to promote the same outcome. In
practice they may be mutually antagonistic. Furthermore, there is good reason to think that the type
of incentive schemes often promoted by agency theorists for use within corporations have
considerable potential to undermine moral motivation. Far from intensifying work effort, the
external incentive scheme may simply communicate the message that management does not trust
workers. One need only recall the way that workers typically respond to sharp incentives such as
piece rates, along with the monitoring systems that are required in order to implement them, to see
the consequences this may have.

The general problem is that agency theory has a completely top-down focus when it comes
analyzing relationships within the firm. Sanctions flow from the principal, who occupies a higher
rank in the organizational hierarchy, down toward the agent, who occupies a subordinate role. It is a
purely unilateral and one-sided relationship. Thus the framing effect of agency theory tends to
encourage essentially Taylorian management practices. There is nothing in the agency perspective,
for instance, that discourages the principal from acting opportunistically with respect to the agent, or
even speaks to this problem.54

Moral relations, on the other hand, are based upon trust, and are therefore typically secured
through some form of reciprocity. Managers cannot dictate that employees exhibit trust, they must
53John Rawls, Political Liberalism (New York: Columbia University Press, 1993), pp. 48-54.
54Dees, Principals, Agents, and Ethics, p. 49.

1
4

work to cultivate it. The standard way of doing this is to exhibit loyalty and trustworthiness in ones
own conduct. Thus moral incentives usually develop within relations that are mutual and two-sided.
These can be extraordinary difficult to cultivate in an environment in which one party also has
unilateral and arbitrary control over the power to punish and reward the other. Thus an organization
that seeks to cultivate trust and loyalty will often go out of its way to downplay its hierarchical
structure, along with the potential for unilateral action that this creates. Thus the type of incentives
schemes that tend to flow from an agency analysis often create an ethos that is highly antagonistic
to the development of strong bonds of solidarity. (One can see here the substance of Kuliks
complaint, that an overemphasis on performance pay, bonuses, and other sharp incentives at Enron
created an agency culture, that in turn eroded the basis for ethical conduct.)

3; Cryptonormativism. No matter how strenuously agency theorists may insist that theirs is only
a positive theory of the firm, and thus entails no value judgments, the fact is that the
basic approach has as its foundation a normative theory of practical rationality, one which
categorizes certain forms of action as rational and certain other forms as irrational. The
fact that morality (or cooperation) gets consistently categorized within such models as
irrational, and opportunism (or defection) as rational, might easily lead more impressionable
minds to the conclusion that they should learn to ignore moral constraints. This can have two
pernicious consequences. First, in the interests of acting more rationally, individuals may
begin to plan their own behavior in accordance with the dictates of the instrumental model,
and thus begin to act more opportunistically. Second, even if they do not change their own
deliberative processes, they may begin to expect higher levels of opportunistic behavior from
others, and therefore feel justified in engaging in preemptive defection in order to protect
themselves from the anticipated defection of others. Thus Ronald Duska observes that the
instrumental conception of rationality has the potential to become a self-fulfilling
prophecy. If I think humans are always going to be selfish, and cannot help but be so, it
becomes the height of foolishness to sacrifice myself, or to predict their behavior on any
other than selfish grounds.55 Yet the type of I did it to him to prevent him from doing it to
me reasoning that this generates provides another one of the classic techniques of
neutralization used to excuse anti-social behavior.56

There is some evidence to support this concern about instrumental rationality becoming a
self- fulfilling prophecy. It was widely reported, for instance, that one of the only significant
anomalies discovered in experimental trials of the public good game in North America occurred
when the game was played among economics graduate students. There the rate of cooperation fell to
only 20 per cent, whereas it remained over 40 per cent when played by students in other
disciplines.57 In a series of follow-up questions, students were asked whether a concern over
fairness played a role in their decisions. Whereas virtually all noneconomists answered yes, more
than one-third of the economists either refused to answer the question regarding what is fair, or gave
55Duska, Why be a Loyal Agent? A Systemic Ethical Analysis, p. 149. Substitute fickle for selfish and the argument is
quibble-proof.
56See Gresham M. Sykes and David Matza, Techniques of Neutralization: A Theory of Delinquency, American Sociological
Review, 22 (1957): 664-670 at 668.
57Gerald Marwell and Ruth E. Ames, Economists Free Ride, Does Anyone Else? Journal of Public Economics, 15, (1981):
295-310. Results of the study need to be taken with a grain of salt, for a variety of reasons. See Robert H. Frank, Thomas
Gilovich and Dennis T. Regan, Does Studying Economics Inhibit Cooperation? Journal of Economic Perspectives, 7 (1993):
159-171.

1
5

very complex, uncodable responses... Those who did respond were much more likely to say that
little or no contribution was

fair. In addition, the economics graduate students were about half as likely as other
subjects to indicate that they were concerned with fairness in making their decisions.58

This is important because, contrary to the widespread conviction that the willingness to act
morally is primarily dependent upon ethical character, which in turn is instilled through childhood
socialization, empirical studies have generated strong support for the contention that the willingness
to act morally is in fact highly situational, and that individuals rely to an exceptional degree upon
social cues in their immediate environment in order to determine what to do. 59 Thus it would be no
surprise to discover that a social environment in which the dominant assumption is that its every
man for himself, is one that would not only encourage unethical behavior, but could become
positively criminogenic.

3. Agency theory as critical theory

The discussion so far has focused upon the mischief that can be caused by an overly-literal
use of agency theory as a tool for understanding the relations between individuals within a firm. The
problems stem from the model of rational action underlying agency theory, which is not normatively
neutral, but results rather in a selective emphasis upon the consequentialist dimension of practical
rationality, while ignoring the role of deontic constraint. Thus the use of agency theory as the
methodological foundation of a positive theory of the firm tends to produce a highly distorted image
of how these organizations function, which can in turn have undesirable effects upon behavior if
naively adopted as an accurate account of reality.

This is, however, not the only way to use agency theory. There is a long-standing tradition in
political philosophy, dating back most obviously to Thomas Hobbes, that uses an instrumental model
of

rationality as the basis for the development of a normative theory.60 Theorists working in this
tradition, rather than asserting that individuals always act in a self-interested manner, instead merely
pose the question, what if individuals always acted in a purely self-interested manner? The
instrumental model is then used as a foundation for the a dystopian state of nature thought58Marwell and Ames, Economists Free Ride, Does Anyone Else? p. 309.
59John Doris, Lack of Character (Cambridge: Cambridge University Press, 2002).
60See Joseph Heath, Rational Choice as Critical Theory, Philosophy and Social Criticism, 22 (1996):
43-62.

1
6

experiment, which characterizes the condition that society would be in if individuals failed to respect
any internal or moral constraints in the way that they pursue their objectives. It is not difficult to
show that, under such conditions, individuals would become embroiled in insuperable collective
action problems. With a little more work (andpace Hobbes), it is possible to show that no system of
purely external incentives can be created that will resolve these problems. 61 Thus a general case can
be made for the claim that individuals should adopt some form of internal constraint, as the best way
of avoiding a life that is solitary, poor, nasty, brutish and short.62

In this tradition of thought, the instrumental conception of rationality is used to construct a

cautionary tale. It allows one to state with a great degree of precision what would happen in the
absence of morality and other systems of deontic constraint. Agency theory can be (and has been)
used in exactly the same way. Thus many business ethicists have drawn upon its results - especially
the limitative results, which show how ubiquitous moral hazard problems would be, and how
difficult the design of effective incentive schemes would be, in the absence of moral constraint - in
order to show that corporations could not even begin to function in the absence of significant moral
constraint on the part of employees or managers. Noreen has developed this insight into a powerful
rebuttal of the standard invisible hand critique of business ethics, which claims that marketplace
competition

renders the constraints of morality otiose.63 As Noreen puts it, agency theory can be used to provide
a series of instructive parables that illustrate the adverse consequences on social and economic
systems of unconstrained opportunistic behavior,64 and can therefore be used as a way of building
the case for ethical conduct in business relations.

According to this perspective, individuals are capable of acting opportunistically, but are also
capable of exhibiting restraint. The extent to which they do either is very much dependent upon
circumstance, institutional context, and background culture. Agency theory offers a characterization
of the dystopian extreme, in which opportunistic conduct is rampant. This provides not only a good
reason for wanting to ensure that greater moral restraint is exercised (viz. to achieve a reduction in
agency costs), it also provides a good explanation for the competitive advantage certain firms are
able to derive from an organizational culture that promotes such restraint. Francis Fukuyama, for
example, has developed this analysis as a way of explaining the competitive advantage that familyowned firms often enjoy in the incubation stage of corporate development. 65 The fact that family
members are able to draw upon preexisting trust relations allows them avoid all sorts of contracting
and agency costs that rival firms must incur. This explains why social capital - the degree to
which communities share norms and values and are able to subordinate individual interests to those

61See David Braybrooke, The Insoluble Problem of the Social Contract, Dialogue, 15 (1976): 3-37.
62David Gauthier, Morals by Agreement (Oxford: Clarendon, 1986).
63Noreen, The Economics of Ethics: A New Perspective on Agency Theory. For an example of the invisible hand critique,
see David Gauthier, No Need for Morality: The Case of the Perfectly Competitive Market, Philosophical Exchange, 3
(1982): 41-54.
64Noreen, The Economics of Ethics: A New Perspective on Agency Theory, p. 360.
65Francis Fukuyama, Trust, (London: Penguin, 1995), pp. 74-80.

1
7

of larger groups66 - is a form of capital. It is precisely because it can be drawn upon by individuals
in order to avoid agency costs in their organizations, both by reducing agency losses directly and by
reducing the need for costly monitoring.

A critical agency perspective is also able to explain quite clearly why, as production becomes
more knowledge-intensive, there is a need to shift away from more Taylorian management strategies,
in favor of a focus upon organizational culture, team-building and shared values. It is because
agency problems are caused, fundamentally, by information asymmetries. As production becomes
more knowledge-intensive, the potential for such problems increases, the difficulty of creating
effective external incentives schemes is compounded, and the probability of such schemes
backfiring increases. Thus firms come to rely more and more upon internal incentives to secure the
voluntary cooperation of their workers. This in turn requires treating them less like cogs, and more
like partners in the production process. (In this respect, critical agency analysis vindicates several of
the fundamental intuitions underlying Peter Druckers analysis of management as a liberal art.67)

Allen Buchanan has taken this style of analysis one step further, arguing that agency theory
not only provides a good argument for business ethics in general, but that the analysis of agency
risks provides the key to understanding many of the real-world moral codes that already (implicitly
or explicitly) structure activities within bureaucratic organizations. His analysis derives important
features of the ethics of bureaucratic organizations from an understanding of what bureaucratic
organizations are like, in particular, from an understanding of what kinds of agency-risks arise within
them.68 Agency theory tells us where the major stress lines are within these organizations, where
cracks are most likely to appear. The ethical code of the organization is then analyzed as the glue that
(to a greater or lesser degree of success) holds things together. Thus in Buchanans view, agency
analysis provides greater theoretical purchase upon these codes, helping business ethicists gain a
greater appreciation of their deep structure.

Buchanan proposes an ingenious analysis, in which he distinguishes between first-order

and second-order agency risks. The former reflect the possibility of actions that imposes costs
upon the principal and benefit the agent directly. The latter involve actions that impose costs upon
the principal, yet benefit the agent only indirectly, insofar as they make it more difficult for the
principal to eliminate first-order agency risks, and thus allow the agent to continue some course of
action that is a source of direct benefit. For example, while shirking would be a first-order agency
problem, employees may also take actions aimed at frustrating a monitoring system that has been
instituted in order to control shirking. In effect, they act to preserve the information asymmetry that
creates the first-order moral hazard problem. Insofar as this is costly to the organization, it is a
second-order agency problem. In this way, Buchanan is able to explain why individuals in
bureaucratic organizations develop a moral allegiance, not just to meritocratic or work ethic
principles, but also to procedures that ensure accountability, due procedure and preservation of the
66Fukuyama, Trust, p. 10. See also James S. Coleman, Social Capital in the Creation of Human Capital, American Journal
of Sociology, 94 (1988): S95-S120.
67Peter Drucker, Management as Social Function and Liberal Art, in The Essential Drucker (New York:
HarperBusiness, 2001): 3-13.
68Buchanan, Toward a Theory of the Ethics of Bureaucratic Organizations, p. 422.

1
8

chain of command.69

The agency perspective is similarly useful when it comes to analyzing the obligations of
senior managers. While agency theory itself is neutral with respect to the doctrine of shareholder
primacy - indeed, the vocabulary is sufficiently broad that the manager can be characterized as the
agent of any of the firms patrons - it is able to explain why, in a standard business corporation,
managers bear special fiduciary obligations towards shareholders.70 Unlike other patron groups,
whose interests are protected by contract, shareholders are residual claimants, with only formal
control of the firms board of directors as a mechanism for ensuring that their interests are respected.
The potential agency risks are simply much greater in the relationship between management and the
firms owners. Thus the agency perspective is able to explain why courts essentially impose a
fiduciary obligation upon senior managers to advance the interests of the owners of the firm (and
why they limit the ability of the parties to contract around this obligation). Agency analysis is also
able to demonstrate quite clearly the ineffectiveness of equity-based compensation arrangements,
such as stock options, as a way of creating an external alignment of managerial and shareholder
interests. Furthermore, it is able to show how the movement of the stock price, combined with the
takeover threat, is an extremely blunt instrument for disciplining management. 71 This in turn helps to
make the case for the claim that moral restraint on the part of managers - a genuine commitment to
serving the shareholder - is an essential element in the proper functioning of the private enterprise
system.

When used in this way, far from being a contributing factor to the recent spate of corporate
scandals, agency theory proves to be an invaluable tool in understanding what went wrong at these
firms. After all, the frauds in question occurred at precisely the points that agency theory identifies as
central fault lines.72 To draw an analogy, consider what an agency analysis of the professional role of
the doctor would look like. There is no question that doctors should exercise moral restraint in their
dealings with colleagues, other medical professionals (nurses, techs, etc.), patients, and their
families.

At the same time, an agency analysis is able to identify patients as the class of individuals who are
uniquely vulnerable to exploitation in their relationship with the doctor (first and foremost because
of the information asymmetries that exist between the two, but also because of the psychological
effects of ill health). Thus the case can be made for a special fiduciary obligation on the part of the
doctor toward the patient (in the same way that a case can be made for a fiduciary obligation
between the manager and the owners of a firm). From this perspective, it would be unsurprising to
discover that most of the internal disciplinary proceedings that occur within physician associations
involve abuse of patients (and not, for example, colleagues). 73 In the same way, would be
69Buchanan, Toward a Theory of the Ethics of Bureaucratic Organizations, p. 431.
70Alexei Marcoux, "A Fiduciary Argument Against Stakeholder Theory," Business Ethics Quarterly,13 (2003): 1-24.
71 See Gary Miller, Managerial Dilemmas (Cambridge: Cambridge University Press, 1992), pp. 171-176. Also Lucian Arye
Bebchuck, Jesse M. Fried and David I. Walker, Managerial Power and Rent Extraction in the Design of Executive
Compensation, University of Chicago Law Review, 69 (2002): 751-846.
72See Lucian Arye Bebchuk and Jesse M. Fried, Excecutive Compensation as an Agency Problem, Journal of Economic
Perspectives, 17 (2003): 71-92.
73Ref. Bulletin of College of Physicians and Surgeons of Canada.

1
9

unsurprising to discover that the major ethics scandals in the corporate world involved an abuse of
shareholders by management.

Of course, it is important not to think that the moral codes of bureaucratic organizations
serve no purpose other than the reduction of agency costs. This would imply a system of moral
constraint containing purely bottom-up obligations, a structure that precisely tracked the
organizational hierarchy of the firm. (The principles that Buchanan outlines, for instance, have this
structure - they all specify what subordinates owe to their superiors. In part for this reason, Buchanan
is it pains to emphasize that his is not a complete conception of business ethics.) An entirely bottom
up moral code would be in tension with the usual system of reciprocity upon which moral
obligations depend. As a result, agency theory may serve a useful purpose in telling us where ethics
is most needed within organizations, but morality has its own logic, and so we many not be able to
develop an ethics code that is custom-tailored to resolve precisely this set of agency problems. There
will often be a quid pro quo, such that ethical conduct can only be elicited from the agent in one
domain if the principal is willing to accept moral constraint in some other domain, where he or she
might have preferred to exercise the freedom to act strategically.

In this context, it is worth emphasizing that the ability of an ethics code to resolve agency
problems can also be quite limited. In particular, while the presence of external sanctions can have
the effect of undermining moral motivation, the absence of external sanctions can also have the
effect of unraveling cooperation (as those who have acted cooperatively in the past become less
willing to do so, when they see others defecting with impunity). Thus it is important, when applying
the critical agency perspective, to keep in mind the limits and the instability of voluntary cooperative
action. The sort of ethics codes typically recommended from the critical agency perspective will
usually not be incentive- compatible, yet that does not give the critical agency theorist license to
ignore the incentives that agents face altogether (or to imagine that ethics is some sort of magic
bullet for resolving agency problems). The fact that a particular institutional arrangement generates
an agency problem in theory is not in itself a problem; but if it has been shown that the arrangement
generates the problem in practice, and the parties seem resistant to moral suasion, then it is time to
start thinking about legal and institutional remedies. Thus when doing applied ethics, it is important
to keep in mind what Rawls calls the strains of commitment. 74 A lot of problems would go away if
people only behaved more ethically, but the fact is, people often dont behave all that ethically. Thus
merely urging more ethical behavior upon them, beyond a certain point, no longer counts as offering
a solution.

4;

Agents for the greedy?

One might think that this sort of critical use of agency theory would be uncontroversial
among business ethicists. Yet that would also be mistaken. There is another prominent line of
objection to agency theory, this time one that condemns both positive and normative uses of the
74John Rawls, A Theory of Justice, 2nd edn. rev (Cambridge, MA: Harvard University Press, 1999), pp. 154-5.

2
0

theory. It is centered upon the claim that agency relationships, even fiduciary relationships, cannot
serve as a genuine source of moral obligation. Under the best of circumstances, they serve only to
transmit moral obligations, from principals to agents. More often, agency relationships are used as
an excuse for unethical conduct, as agents seek to avoid responsibility by claiming that they are
merely following orders or serving the client. From this perspective, agency theory is nothing
but a giant distraction, a way of passing the buck when it comes to confronting the problem of
unethical behavior in business. Either the agents action is ethical, in which case the agency
relationship has nothing to do with it and the source must be traced back to some obligation imposed
upon the principal, or it is unethical, and the agency relationship serves only to obscure that fact, by
suggesting that it was done out of loyalty or obligation to the principal. In both cases, the agency
relationship has nothing to do with the moral obligations that individuals are subject to, and so
business ethicists gain nothing by focusing upon it.

This view does have someprimafacie plausibility. It is a well-known feature of conventional

morality that promising to help a friend commit a crime does not generate a moral obligation on
ones part to commit that crime (such that if ones friend were to fall ill on the eve of the bank
robbery, one might be morally obliged to carry it out anyway on his behalf). To allow this would be
to permit the unlimited laundering of unethical acts into ethical ones. Yet many people seem to
believe that professional roles do permit laundering of this sort. Thus, for example, what might
ordinarily be regarded as lying is sometimes presented, not just as permissible, but as morally
obligatory, when done by a lawyer who is seeking to advance the interests of a client. 75 Arthur
Applbaum draws out the absurd consequences of such a view of role obligations by developing a
profile of Sanson, the executioner of Paris, who carried out his duties with consummate
professionalism throughout the final years of the ancien regime, the French Revolution, the Terror
and the Thermidor. He remained above the fray throughout, by insisting that he was merely a loyal
agent, carrying out legal executions, and was thus not to be held responsible for any of the excesses
committed by one or another of the various principals he had served.

More generally, Applbaum develops a thought-experiment involving two societies, Badland

and Roland. Badland is essentially a Hobbesian state of nature, in which each individual pursues his
or her self-interest in a purely instrumental fashion, and thus no one avoids harming another unless
there are penalties discouraging such harm, and all craftily engage in manipulation and deception if
doing so will advance their ends.76 In Roland, by contrast, people have the same motivations, but
do not pursue their own interests. Rather, each appoints a trustee who pledges to advance the
trustors interests through a blind trust, and each trustor is also a trustee. As a result, in Roland
exactly the same conflicts are fought, the same manipulations occur, the same harms inflicted, but
each actor is acting as a faithful professional in fulfillment of obligations to a client. 77 In what
sense, Applbaum then asks, is Roland any better than Badland?

Many business ethicists have seen the relationship between managers and shareholders as
essentially equivalent to the relationship between trustees and trustors in Roland. Alex Michalos, for
75See Arthur Isak Applbaum, Ethics for Adversaries (Princeton: Princeton University Press, 1999), p. 77.
76Applbaum, Ethics for Adversaries, p. 7.
77Applbaum, Ethics for Adversaries, p. 8.

2
1

instance, in his critique of the loyal agents argument, attributes the following view to theorists
(like Milton Friedman) who view the obligation of managers toward shareholders as paramount: As
a loyal agent of some principal, I ought to serve his interests as he would serve them himself... He
would serve

his own interests in a thoroughly egoistic way. Therefore, as a loyal agent of this principal, I ought to
operate in a thoroughly egoistic way on his behalf. 78 Michalos goes on to criticize this argument,
claiming that it adds up to little more than an attempt to launder egoism into altruism.

More polemically, Lisa Newton has argued that, from the perspective of agency theory, the
entirety of corporate enterprise seems... to be dedicated to the enrichment of the rich, to satisfy the
greed of the truly greedy.79 Furthermore, it follows for agency theory that there can be no such
thing as corporate responsibility for community welfare, for the community figures nowhere in the
principal- agent relationships.80 Thus she refers to the moral framework encouraged by agency
analysis as theory-compelled irresponsibility.

There are two separate issues in this agents for the greedy critique, which need to be
disentangled:

1; The nemo datproblem. In discussing principal-agent relations, Kenneth Goodpaster refers to

the Latin proverb, nemo dat quod non habet, or nobody gives what he doesnt have. He
uses this to draw attention to the fact that agency relationships are unable to create moral
permissions where previously none existed. Principals cannot (ethically) hire someone to do
on their behalf what they could not (ethically) do themselves.81 In a similar vein, Richard
DeGeorge takes pains to emphasize that, acting for another does not give one ethical
license, and that all persons are ethically responsible for their actions, whether performed
under command or performed on behalf of another.82 Yet since the agency relationship
cannot be a source of moral permissions, it is then claimed, the fact that managers act, in

some sense, as agents of shareholders, is simply lacking in moral significance.

Yet while these observations are correct, the conclusion does not follow. There is in fact
78Alex C. Michalos, A Pragmatic Approach to Business Ethics (Thousand Oaks: Sage, 1995), p. 45 [format altered].
79Lisa Newton, Agents for the Truly Greedy, in Bowie and Freeman, eds. Ethics and Agency Theory, p. 100.
80Newton, Agents for the Truly Greedy, p. 101.
81Kenneth E. Goodpaster, Business Ethics and Stakeholder Analysis, Business Ethics Quarterly, 1 (1991): 53-73, at 68
[slightly modified, text of the article is corrupt].
82Richard T. DeGeorge, Agency Theory and the Ethics of Agency, in Bowie and Freeman, eds. Ethics and Agency Theory,
pp. 65-66.

2
2

some danger of equivocation in the talk about ethical and unethical conduct. With respect to
agents, it is important to distinguish the deontic modalities of permission and obligation. Critics of
the agency perspective are perfectly correct in noting that agency relations cannot create
permissions. This is in fact why theorists who are heavily influenced by the agency perspective, such
as Buchanan, are at pains to specify that the moral obligations of managers are to advance the
legitimate interests of shareholders (not any-old-interests).83 Even Friedman qualifies his defense of
profit-maximization (in an unfortunately glib manner), with the stipulation that shareholders will
generally want to make as much money as possible while conforming to the basic rules of the
society, both those embodied in law and those embodied in ethical custom. 84 Thus no one is
committing the elementary error of believing that agency relations can turn impermissible conduct
into permissible conduct (or wrong into right).

What critics of the agency perspective generally fail to note is that agency relations can serve
as a genuine source of moral obligation in one important sense - agency relations can transform
actions that are merely permissible for the principal into ones that are obligatory for the agent. This
is in fact Applbaums final observation in Ethics for Adversaries. In response to the (rhetorical)
question, Why take professional roles seriously, from the moral point of view? he replies: Though
roles ordinarily cannot permit what is forbidden, they can require what is permitted. 85 For example,
a person who is accused of a crime, even though he may have done it, is not obliged to plead guilty,
but rather is permitted to mount a defense (legally of course, but perhaps also morally, in cases
where the prosecution is seeking an unreasonably harsh sentence). Yet mounting a defense is, for the
accused, merely the exercise of a permission (as witnessed by the fact that he is entitled, at any
point, to change his mind and enter a guilty plea). For any attorney that he employs, on the other
hand, the exercise of this permission generates an obligation to mount that defense.

The relationship between the principal and the agent can be summed up in the following set
of inferences. Consider the case where the principal instructs the agent to perform some action, a,
that falls within the scope of a principal-agent relationship:

1; If a is obligatory for the principal then a is obligatory for the agent.

2; If a is permissible for the principal, then a is obligatory for the agent.

3; If a is forbidden for the principal, then a is forbidden for the agent.

83Buchanan, Toward a Theory of the Ethics of Bureaucratic Organizations, pp. 422-23.
84Milton Friedman, The Social Responsibility of Business is the Increase its Profits, New York Times Magazine, (Sept.
13, 1970).
85Applbaum, Ethics for Adversaries, p. 259.

2
3

Thus, from the standpoint of business ethics, insofar as shareholders are permitted to claim
the residual earnings of the firm, and are permitted to seek maximization of those earnings, then
managers are obliged to serve them loyally in this regard. This is morally salient, because the agency
relationship creates that moral obligation, by transforming a permission into an obligation. Thus the
nemo dat principle is misleading, and Goodpasters observation does not undermine the significance
of agency analysis. Agency relationships are not merely a distraction; they represent a genuine
source of moral obligation.

2. Greed is not good. The second major issue concerns the moral status of shareholder interests, and
the intuition that there is something morally dubious about the desire of shareholders to make as
much money as possible (as Friedman so gracefully put it). Of course, one need not show that there
is anything laudable about the desire for profit in order to demonstrate the importance of agency
analysis - one need only demonstrate that it is morally permissible. This is a much lighter burden of
proof, a fact that is sometimes obscured by theorists like Newton, who use abstract terms of
condemnation such as greed. Certain actions may not be morally praiseworthy, but they are not, by
virtue of that fact, morally impermissible.

Nevertheless, there is a persistent intuition among ethicists (and the broader public), that
making money cant be the foundation of business ethics, regardless of whether that money is made
for oneself or for someone else. Making money just doesnt seem ethical enough. 86 Naturally, there
are those who disagree. Winston Churchill put it rather pithily, when he declared that it is a
Socialist idea that making profits is a vice: I consider the real vice is making losses. Kenneth
Arrow, in his reflections on business ethics, adopted a similar point of view. Profit, he said, really
represents the net contribution that the firm makes to the social good, and profits should therefore be
made as large as possible.87 Yet this view has never succeeded in carrying general conviction. (One
need only consider the demonization of the profit motive that one finds in recent films, like The
Corporation.)

There is a large constituency of business ethicists who are similarly skeptical about the profit
motive. Duska, for instance, suggests that profit is merely a way of motivating managers, but that
the real purpose of the business enterprise lies elsewhere. The actual goal of business, he claims, is
to help increase the production of goods and services. He then suggests that it is a mistake to
confuse incentives for good actions with the purpose of those actions. Good grades are incentives for
study, but the purpose of study is primarily to learn, to expand the mind. 88 There are circumstances
in which these two objectives can come apart. Students who engage in too single-minded a pursuit of
grades often fail to expand their minds. Similarly, in a business context, there is potential for
divergence between first, the social purpose of providing goods and services; and second, the
agents altruistic commitment to the principal.89 In this case, it is the true purpose of the business
that should serve as the object of the managers true loyalty, rather than the demands of the principal
86 Goodpaster formulates this intuition quite well in his discussion of the so-called stakeholder paradox. See Goodpaster,
Business Ethics and Stakeholder Analysis, p. 63.
8787Kenneth Arrow, Social Responsibility and Economic Efficiency, Public Policy, 21 (1973): 303-317 at 305.
88Duska, Why be a Loyal Agent? p. 161.
89Duska, Why be a Loyal Agent? p. 161.

2
4

for profit. Thus any

obligations arising out the agency relationship are easily trumped, in Duska's view, by
considerations that pertain to the true purpose of the firm.

There is something right about this, but also something importantly wrong. What critics of
the profit motive typically fail to take into consideration is the fact that the market economy is an
adversarial institution.90 Adversarial institutions achieve the results that they do by playing off
various parties against one another, producing outcomes that correspond only accidentally to those
intended or desired by any of the participants. Consider the adversarial structure of a criminal trial.
The overall goal is to ensure that justice is served (and that conviction of the innocent is
minimized). But this outcome is achieved, not by instructing all of the major parties to pursue the
outcome that they consider to be the most just. On the contrary, the crown (or district) attorney is
expected to exhibit prosecutorial zeal in seeking a conviction, regardless of what she may regard
as the merits of the case. Similarly, the defense attorney is expected to mount a vigorous defense,
independent of any opinions he may have about the defendants guilt or innocence. The thought is
then that these two forces cancel one another out, leaving the judge free (as much as possible) to
assess the case on its true merits. Thus the prosecution and the defense, although they participate in a
system that has as its goal the production of justice, are not obliged to adopt that outcome as their
objective. Indeed, they are often barred from doing so. Imagine a defense attorney with access to
privileged information that confirmed the guilt of her client. It would be the height of professional
misconduct for her to leak this information to the prosecution, or to sabotage her own clients case,
simply because she wanted to ensure that justice was served.

One can see a similar structure in the competitive market. Profit itself is not the goal of the
institution. On the contrary, the point of having a competitive market is to play various profitseeking corporations off against one another (often on both the supply and the demand side) in order
to achieve pressure in the direction of market-clearing prices. This in turn reduces deadweight losses,
and thus ensures that all resources are put to their best employment. Under ideal conditions (i.e. in a
perfectly competitive market), the strategies adopted by the parties completely cancel one another
out - hence the absence of any profit in general equilibrium. (This is also why, in sectors where it is
impossible to organize a competitive market, firms are usually subject to regulatory constraints that
prevent them from adopting profit-maximizing strategies.) Thus John Kay gets things right when he
claims that it is not true that profit is the purpose of the market economy, and the production of
goods and services the means to it: the purpose is the production of goods and services, profit the
means.91 But one cannot infer from this, as Duska does, that the production of goods and services is
therefore that true purpose of the corporation, and that the manager should assign priority to this
objective whenever it conflicts with the interests of the principal. This would be like claiming that
the true purpose of the defense attorney is to see that justice is served. Both managers and
attorneys should respect certain moral constraints in choosing the strategies that they will employ.
But there is an enormous difference between respecting certain moral constraints while still
advancing the objectives of the principal, and substituting some other set of objectives for the
90[self-identifying note removed]
91John Kay, The Truth About Markets (London: Penguin, 2003), p. 351.

2
5

principals, on the grounds that ones own moral reasoning leads one to assign them greater
priority.92

It is the adversarial nature of markets that explains the whiff of immorality that has always
accompanied the profit motive. Those who demonize corporations for pursuing profits have much in
common with those who denounce criminal defense lawyers for defending rapists and murderers.
In the case of doctors and patients, its usually obvious that the health of the patient is a good thing,
and therefore that the doctor is doing a good thing in helping the patient to secure it. In the case of
managers and shareholders, on the other hand, it is not obvious that profit is a good thing, just as
with defense

attorneys and those who have been accused of a crime, it is not obvious that acquittal is a good thing.

In order to see the full justification for the objectives that the agent is pursuing, it is necessary to
situate them within the broader context of the institution within which they are pursued, and in
particular, to take into consideration the countervailing forces that are set up within that same
institution, which serve as a check upon the ability of any one party to achieve its objectives.

Thus the normative critique of the agency perspective is based upon a set of conceptual
confusions. Of course, there is something genuinely counterintuitive about that idea that managers
might be morally obliged to maximize the profits of shareholders (or act as agents for the greedy).
It sounds wrong when one first hears it. But upon closer examination, it turns out to be perfectly
defensible. So while the obligations that managers have toward the owners of a firm will most
certainly not add up to all of business ethics, they will certainly form the core of any defensible
conception of business ethics. When seen in this light, the big ethical question becomes, not whether
managers should advance the interests of the owners of the firm, but rather how far managers should
go in advancing the interests of the firms owners. (In the same way, the major question in legal
ethics is not whether the lawyer should serve the client, but rather how far the lawyer should go in
advancing the interests of the client.) In the case of markets, this question admits of a very
straightforward answer. Economists use the term market failure to refer to situations in which the
competitive pursuit of profit by individual firms fails to achieve the purpose of the market economy,
viz. the efficient production of goods and services. This provides a set of relatively straightforward
guidelines for distinguishing permissible from impermissible profit-maximizing strategies. Taking
advantage of a market failure in order to increase profits (e.g. through the creation of negative
externalities, exploitation of information asymmetries, exercise of market power, etc.) are
impermissible.93 Anything else is fair game.

92For further analysis, see [self-identifying note].

93This is the core of Arrows view (Social Responsibility and Economic Efficiency,), although he focuses only on the twin
cases of pollution and information asymmetry.

2
6

5;

Conclusion

The preceding discussion has examined two very different strategies for employing agency
theory - the positive and the normative - and two very different sorts of objections that have been
raised by business ethicists. The use of agency theory brings to the fore two sets of ideas that
ethicists have traditionally been very uncomfortable with, first, the economic model of rational
action, and second, the doctrine of shareholder primacy and the status of profit. With regard to the
first, I have suggested that business ethicists have been fully justified in their reservations. The
economic model is based upon an inadequate conception of rational action, precisely because it
classifies an important category of moral action as irrational. Indeed, it classifies all genuine rulefollowing as irrational, and is therefore unsuitable for use as a general theory of rational action.
Sophisticated practitioners of agency theory are familiar with these limitations, but a large number of
enthusiasts are not. Thus agency theory can serve as a source of considerable inadvertent mischief
when treated as an accurate representation of reality. I have therefore encouraged a critical use of
agency theory, in which principal-agent analysis is used to provide, not a model of how firms
actually work, but rather a set of instructive parables, allowing us to see more clearly what the
world of business would be like in the absence of business ethics.

With regard to the doctrine of shareholder primacy, and the extent to which agency theory
encourages this perspective, I have tried to emphasize that there is no simply connection between the
two. In this context, it is perhaps worth noting that R. Edward Freeman used principal-agent
vocabulary as a way of articulating his commitment to a stakeholder approach to business ethics.
Indeed, he even presented an Agency Principle as an element of his Doctrine of Fair Contracts,
intended to guide the management of stakeholder-focused firms. 94 Thus agency theory is neutral in
this regard.

Nevertheless, when employed cautiously, with due attention to the institutional context in which the

firm operates, it is possible to use agency theory as the basis for a plausible shareholder-focused
conception of business ethics. Agency theory can be used to show how the shareholder is in a
uniquely vulnerable position with respect to the manager, and therefore why a fiduciary relation is
justifiable in this case. So while a commitment to agency analysis neither presupposes nor entails a
commitment to the doctrine of shareholder primacy, the gain in conceptual clarity afforded by the
agency perspective does provide a powerful source of arguments in favor of that doctrine.

94R. Edward Freeman, A Stakeholder Theory of the Modern Corporation, in Clarkson, ed. The Corporation and its
Stakeholders, p. 136.

2
7