Article Summaries

Agency Theory: An Assessment and ReviewAuthor(s): Kathleen M. Eisenhardt
Reviewed work(s):
Source: The Academy of Management Review, Vol. 14, No. 1 (Jan., 1989), pp. 57-74
Published by: Academy of Management
Stable URL: http://www.jstor.org/stable/258191 .
Accessed: 24/08/2012 14:51
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
.
Academy of Management is collaborating with JSTOR to digitize, preserve and extend access to The Academy
of Management Review.
http://www.jstor.org
? Academy of Management Review, 1989, Vol. 14, No. 1, 57-74.
Agency
Theory:
and
An
Assessment
Review
KATHLEEN
M. EISENHARDT
Stanford University
Agency theory is an important, yet controversial, theory. This paper
reviews agency theory, its contributions to organization theory, and
the extant empirical work and develops testable propositions. The
conclusions are that agency theory (a) offers unique insight into information systems, outcome uncertainty, incentives, and risk and (b)
is an empirically valid perspective, particularly when coupled with
complementary perspectives. The principal recommendation is to incorporate an agency perspective in studies of the many problems
having a cooperative structure.
One day Deng Xiaoping decided to take his
grandson to visit Mao. “Call me granduncle,”
Mao offered warmly. “Oh, I certainly couldn’t
do that, Chairman Mao,” the awe-struck child
replied. “Why don’t you give him an apple?”
suggested Deng. No sooner had Mao done so
than the boy happily chirped, “Oh thank you,
Granduncle.” “You see,” said Deng, “what incentives can achieve.” (“Capitalism,”1984, p.
62)
purposes of this paper are to describe agency
theory and to indicate ways in which organizational researchers can use its insights. The paper is organized around four questions that are
germane to organizational research. The first
asks the deceptively simple question, What is
agency theory? Often, the technical style, mathematics, and tautological reasoning of the
agency literature can obscure the theory. Moreover, the agency literature is split into two
camps (Jensen, 1983), leading to differences in
interpretation. For example, Barney and Ouchi
(1986) argued that agency theory emphasizes
how capital markets can affect the firm,
whereas other authors made no reference to
capital markets at all (Anderson, 1985; Demski &
Feltham, 1978; Eccles, 1985; Eisenhardt, 1985).
The second question is, What does agency
theory contribute to organizational theory? Proponents such as Ross (1973, p. 134) argued that
“examples of agency are universal.” Yet other
scholars such as Perrow (1986) claimed that
agency theory addresses no clear problems,
and Hirsch and Friedman (1986) called it excessively narrow, focusing only on stock price. For
economists, long accustomed to treating the or-
Agency theory has been used by scholars in
accounting (e.g., Demski & Feltham, 1978), economics (e.g., Spence & Zeckhauser, 1971), finance (e.g., Fama, 1980), marketing (e.g., Basu,
Lal, Srinivasan, & Staelin, 1985), political science (e.g., Mitnick, 1986), organizational behavior (e.g., Eisenhardt, 1985, 1988; Kosnik, 1987),
and sociology (e.g., Eccles, 1985; White, 1985).
Yet, it is still surrounded by controversy. Its proponents argue that a revolution is at hand and
that “the foundation for a powerful theory of organizations is being put into place” (Jensen,
1983, p. 324). Its detractors call it trivial, dehumanizing, and even “dangerous” (Perrow, 1986,
p. 235).
Which is it: grand theory or great sham? The
57
vision of labor (Jensen & Meckling, 1976; Ross,
1973). Specifically, agency theory is directed at
the ubiquitous agency relationship, in which
one party (the principal) delegates work to another (the agent), who performs that work.
Agency theory attempts to describe this relationship using the metaphor of a contract (Jensen &
Meckling, 1976).
Agency theory is concerned with resolving
two problems that can occur in agency relationships. The first is the agency problem that arises
when (a) the desires or goals of the principal and
agent conflict and (b) it is difficult or expensive
for the principal to verify what the agent is actually doing. The problem here is that the principal cannot verify that the agent has behaved
appropriately. The second is the problem of risk
sharing that arises when the principal and
agent have different attitudes toward risk. The
problem here is that the principal and the agent
may prefer different actions because of the different risk preferences.
Because the unit of analysis is the contract
governing the relationship between the principal and the agent, the focus of the theory is on
determining the most efficient contract governing the principal-agent relationship given assumptions about people (e.g., self-interest,
bounded rationality, risk aversion), organizations (e.g., goal conflict among members), and
information (e.g., information is a commodity
which can be purchased). Specifically, the
question becomes, Is a behavior-oriented contract (e.g., salaries, hierarchical governance)
more efficient than an outcome-oriented contract (e.g., commissions, stock options, transfer
of property rights, market governance)? An overview of agency theory is given in Table 1.
The agency structure is applicable in a variety
of settings, ranging from macrolevel issues such
as regulatory policy to microlevel dyad phenomena such as blame, impression management, lying, and other expressions of selfinterest. Most frequently, agency theory has
been applied to organizational phenomena
ganization as a “black box” in the theory of the
firm, agency theory may be revolutionary. Yet,
for organizational scholars the worth of agency
theory is not so obvious.
The third question is, Is agency theory empirically valid? The power of the empirical research
on agency theory to explain organizational phenomena is important to assess, particularly in
light of the criticism that agency theory is
“hardly subject to empirical test since it rarely
tries to explain actual events” (Perrow, 1986, p.
224). Perrow (1986) also criticized the theory for
being unrealistically one-sided because of its
neglect of potential exploitation of workers.
The final question is, What topics and contexts
are fruitful for organizational researchers who
use agency theory? Identifying how useful
agency theory can be to organizational scholars
requires understanding the situations in which
the agency perspective can provide theoretical
leverage.
The principal contributions of the paper are to
present testable propositions, identify contributions of the theory to organizational thinking,
and evaluate the extant empirical literature. The
overall conclusion is that agency theory is a useful addition to organizational
theory. The
agency theory ideas on risk, outcome uncertainty, incentives, and information systems are
novel contributions to organizational thinking,
and the empirical evidence is supportive of the
theory, particularly when coupled with complementary theoretical perspectives.
Origins of Agency Theory
During the 1960s and early 1970s, economists
explored risk sharing among individuals or
groups (e.g., Arrow, 1971; Wilson, 1968). This
literature described the risk-sharing problem as
one that arises when cooperating parties have
different attitudes toward risk. Agency theory
broadened this risk-sharing literature to include
the so-called agency problem that occurs when
cooperating parties have different goals and di-
58
Table 1
Agency Theory Overview
Key idea
Principal-agent relationships should
reflect efficient organization of
information and risk-bearing costs
Unit of
analysis
Contract between principal and agent
Human
assumptions
Self-interest
Bounded rationality
Risk aversion
Organizational
assumptions
Partial goal conflict among participants
Efficiency as the effectiveness criterion
Information asymmetry between principal
and agent
Information
assumption
Information as a purchasable commodity
Contracting
problems
Agency (moral hazard and adverse
selection)
Risk sharing
Problem
domain
Relationships in which the principal and
agent have partly differing goals and
risk preferences (e.g., compensation,
regulation, leadership, impression
management, whistle-blowing, vertical
integration, transfer pricing)
positivist and principal-agent (Jensen, 1983). The
two streams share a common unit of analysis:
the contract between the principal and the
agent. They also share common assumptions
about people, organizations, and information.
However, they differ in their mathematical rigor,
dependent variable, and style.
Positivist Agency Theory
Positivist researchers have focused on identifying situations in which the principal and agent
are likely to have conflicting goals and then describing the governance mechanisms that limit
the agent’s self-serving behavior. Positivist research is less mathematical than principalagent research. Also, positivist researchers
have focused almost exclusively on the special
case of the principal-agent relationship between
owners and managers of large, public corporations (Berle & Means, 1932).
Three articles have been particularly influential. Jensen and Meckling (1976) explored the
ownership structure of the corporation, including how equity ownership by managers aligns
managers’ interests with those of owners. Fama
(1980) discussed the role of efficient capital and
labor markets as information mechanisms that
are used to control the self-serving behavior of
top executives. Fama and Jensen (1983) described the role of the board of directors as an
information system that the stockholders within
large corporations could use to monitor the opportunism of top executives. Jensen and his colleagues (Jensen, 1984; Jensen & Roeback, 1983)
extended these ideas to controversial practices,
such as golden parachutes and corporate raiding.
From a theoretical perspective, the positivist
stream has been most concerned with describing the governance mechanisms that solve the
agency problem. Jensen (1983, p. 326) described
this interest as “why certain contractual relations arise.” Two propositions capture the governance mechanisms which are identified in the
positivist stream. One proposition is that out-
such as compensation (e.g., Conlon & Parks,
1988; Eisenhardt, 1985), acquisition and diversification strategies (e.g., Amihud & Lev, 1981),
board relationships (e.g., Fama & Jensen, 1983;
Kosnik, 1987), ownership and financing structures (e.g., Argawal & Mandelker, 1987; Jensen
& Meckling, 1976), vertical integration (Anderson, 1985; Eccles, 1985), and innovation (Bolton,
1988; Zenger, 1988). Overall, the domain of
agency theory is relationships that mirror the
basic agency structure of a principal and an
agent who are engaged in cooperative behavior, but have differing goals and differing attitudes toward risk.
Agency Theory
From its roots in information economics,
agency theory has developed along two lines:
59
Principal-Agent Research
come-based contracts are effective in curbing
agent opportunism. The argument is that such
contracts coalign the preferences of agents with
those of the principal because the rewards for
both depend on the same actions, and, therefore, the conflicts of self-interest between principal and agent are reduced. For example, Jensen
and Meckling (1976) described how increasing
the firm ownership of the managers decreases
managerial opportunism. In formal terms,
Principal-agent researchers are concerned
with a general theory of the principal-agent relationship, a theory that can be applied to employer-employee, lawyer-client, buyer-supplier,
and other agency relationships (Harris & Raviv,
1978). Characteristic of formal theory, the principal-agent paradigm involves careful specification of assumptions, which are followed by
logical deduction and mathematical proof.
In comparison with the positivist stream, principal-agent theory is abstract and mathematical
and, therefore, less accessible to organizational
scholars. Indeed, the most vocal critics of the
theory (Perrow, 1986; Hirsch et al., 1987) have
focused their attacks primarily on the more
widely known positivist stream. Also, the principal-agent stream has a broader focus and
greater interest in general, theoretical implications. In contrast, the positivist writers have focused almost exclusively on the special case of
the owner/CEO relationship in the large corporation. Finally, principal-agent
research includes many more testable implications.
For organizational scholars, these differences
provide background for understanding criticism
of the theory. However, they are not crucial.
Rather, the important point is that the two
streams are complementary: Positivist theory
identifies various contract alternatives, and principal-agent theory indicates which contract is
the most efficient under varying levels of outcome uncertainty, risk aversion, information,
and other variables described below.
The focus of the principal-agent literature is
on determining the optimal contract, behavior
versus outcome, between the principal and the
agent. The simple model assumes goal conflict
between principal and agent, an easily measured outcome, and an agent who is more risk
averse than the principal. (Note: The argument
behind a more risk averse agent is that agents,
who are unable to diversify their employment,
should be risk averse and principals, who are
Proposition 1: When the contract between the
principal and agent is outcome based, the
agent is more likely to behave in the interests of
the principal.
The second proposition is that information systems also curb agent opportunism. The argument here is that, since information systems inform the principal about what the agent is actually doing, they are likely to curb agent opportunism because the agent will realize that he or
she cannot deceive the principal. For example,
Fama (1980) described the information effects of
efficient capital and labor markets on managerial opportunism, and Fama and Jensen (1983)
described the information role that boards of directors play in controlling managerial behavior.
In formal terms,
Proposition 2: When the principal has information to verify agent behavior, the agent is more
likely to behave in the interests of the principal.
At its best, positivist agency theory can be regarded as enriching economics by offering a
more complex view of organizations (Jensen,
1983). However, it has been criticized by organizational theorists as minimalist (Hirsch,
Michaels, & Friedman, 1987; Perrow, 1986) and
by microeconomists as tautological and lacking
rigor (Jensen, 1983). Nonetheless,
positivist
agency theory has ignited considerable research (Barney & Ouchi, 1986) and popular interest (“Meet Mike,” 1988).
60
such as budgeting systems, reporting procedures, boards of directors, and additional layers
of management. Such investments reveal the
agent’s behavior to the principal, and the situation reverts to the complete information case. In
formal terms,
capable of diversifying
their investments,
should be risk neutral.) The approach of the simple model can be described in terms of cases
(e.g., Demski & Feltham, 1978). The first case, a
simple case of complete information, is when the
principal knows what the agent has done.
Given that the principal is buying the agent’s
behavior, then a contract that is based on behavior is most efficient. An outcome-based contract would needlessly transfer risk to the agent,
who is assumed to be more risk averse than the
principal.
The second case is when the principal does
not know exactly what the agent has done.
Given the self-interest of the agent, the agent
may or may not have behaved as agreed. The
agency problem arises because (a) the principal
and the agent have different goals and (b) the
principal cannot determine if the agent has behaved appropriately. In the formal literature,
two aspects of the agency problem are cited.
Moral hazard refers to lack of effort on the part of
the agent. The argument here is that the agent
may simply not put forth the agreed-upon effort.
That is, the agent is shirking. For example,
moral hazard occurs when a research scientist
works on a personal research project on company time, but the research is so complex that
corporate management cannot detect what the
scientist is actually doing. Adverse selection refers to the misrepresentation of ability by the
agent. The argument here is that the agent may
claim to have certain skills or abilities when he
or she is hired. Adverse selection arises because
the principal cannot completely verify these
skills or abilities either at the time of hiring or
while the agent is working. For example, adverse selection occurs when a research scientist
claims to have experience in a scientific specialty and the employer cannot judge whether
this is the case.
In the case of unobservable behavior (due to
moral hcazard or adverse selection), the principal
has two options. One is to discover the agent’s
behavior by investing in information systems
Proposition 3: Information systems are positively related to behavior-based contracts and
negatively related to outcome-based contracts.
The other option is to contract on the outcomes
of the agent’s behavior. Such an outcome-based
contract motivates behavior by coalignment of
the agent’s preferences with those of the principal, but at the price of transferring risk to the
agent. The issue of risk arises because outcomes
are only partly a function of behaviors. Government policies, economic climate, competitor actions, technological change, and so on, may
cause uncontrollable variations in outcomes.
The resulting outcome uncertainty introduces
not only the inability to preplan, but also risk
that must be borne by someone. When outcome
uncertainty is low, the costs of shifting risk to the
agent are low and outcome-based contracts are
attractive. However, as uncertainty increases, it
becomes increasingly expensive to shift risk despite the motivational benefits of outcome-based
contracts. In formal terms,
Proposition 4: Outcome uncertainty is positively
related to behavior-based contracts and negatively related to outcome-based contracts.
This simple agency model has been described
in varying ways by many authors (e.g., Demski
& Feltham, 1978; Harris & Raviv, 1979; Holmstrom, 1979; Shavell, 1979). However, the heart
of principal-agent theory is the trade-off between (a) the cost of measuring behavior and (b)
the cost of measuring outcomes and transferring
risk to the agent.
A number of extensions to this simple model
are possible. One is to relax the assumption of a
risk-averse agent (e.g., Harris & Raviv, 1979).
Research (MacCrimmon & Wehrung, 1986) indicates that individuals vary widely in their risk
61
a retail sales cashier is much more programmed
than that of a high-technology entrepreneur.
The argument is that the behavior of agents engaged in more programmed jobs is easier to observe and evaluate. Therefore, the more programmed the task, the more attractive are behavior-based contracts because information
about the agent’s behavior is more readily determined. Very programmed tasks readily reveal agent behavior, and the situation reverts to
the complete information case. Thus, retail sales
clerks are more likely to be paid via behaviorbased contracting (e.g., hourly wages), whereas entrepreneurs are more likely to be compensated with outcome-based contracts (e.g., stock
ownership). In formal terms,
attitudes. As the agent becomes increasingly
less risk averse (e.g., a wealthy agent), it becomes more attractive to pass risk to the agent
using an outcome-based contract. Conversely,
as the agent becomes more risk averse, it is increasingly expensive to pass risk to the agent. In
formal terms,
Proposition 5: The risk aversion of the agent is
positively related to behavior-based contracts
and negatively related to outcome-based contracts.
Similarly, as the principal becomes more risk
averse, it is increasingly attractive to pass risk to
the agent. In formal terms,
Proposition 6: The risk aversion of the principal
is negatively related to behavior-based contracts and positively
related to outcomebased contracts.
Proposition 8: Task programmability is positively related to behavior-based contracts and
negatively related to outcome-based contracts.
Another extension is to relax the assumption
of goal conflict between the principal and agent
(e.g., Demski, 1980). This might occur either in a
highly socialized or clan-oriented firm (Ouchi,
1979) or in situations in which self-interest gives
way to selfless behavior (Perrow, 1986). If there
is no goal conflict, the agent will behave as the
principal would like, regardless of whether his
or her behavior is monitored. As goal conflict
decreases, there is a decreasing motivational
imperative for outcome-based contracting, and
the issue reduces to risk-sharing considerations.
Under the assumption of a risk-averse agent,
behavior-based contracts become more attractive. In formal terms,
Another task characteristic is the measurability of the outcome (Anderson, 1985; Eisenhardt,
1985). The simple model assumes that outcomes
are easily measured. However, some tasks require a long time to complete, involve joint or
team effort, or produce soft outcomes. In these
circumstances, outcomes are either difficult to
measure or difficult to measure within a practical amount of time. When outcomes are measured with difficulty, outcome-based contracts
are less attractive. In contrast, when outcomes
are readily measured, outcome-based contracts
are more attractive. In formal terms,
Proposition 9: Outcome measurability is negatively related to behavior-based contracts and
positively related to outcome-based contracts.
Proposition 7: The goal conflict between principal and agent is negatively related to behaviorbased contracts and positively related to outcome-based contracts.
Finally, it seems reasonable that when principals and agents engage in a long-term relationship, it is likely that the principal will learn
about the agent (e.g., Lambert, 1983) and so will
be able to assess behavior more readily. Conversely, in short-term agency relationships, the
information asymmetry between principal and
agent is likely to be greater, thus making out-
Another set of extensions relates to the task performed by the agent. For example, the progammability of the task is likely to influence the ease
of measuring behavior (Eisenhardt, 1985, 1988).
Programmability
is defined as the degree to
which appropriate behavior by the agent can
be specified in advance. For example, the job of
62
asymmetry is linked to the power of lower order
participants (e.g., Pettigrew, 1973). The difference is that in political models goal conflicts are
resolved through bargaining, negotiation, and
coalitions-the
power mechanism of political
science. In agency theory they are resolved
through the coalignment of incentives-the
price mechanism of economics.
Agency theory also is similar to the information processing approaches to contingency theory (Chandler, 1962; Galbraith, 1973; Lawrence
& Lorsch, 1967). Both perspectives are information theories. They assume that individuals are
boundedly rational and that information is distributed asymmetrically throughout the organization. They also are efficiency theories; that is,
they use efficient processing of information as a
criterion for choosing among various organizing
forms (Galbraith, 1973). The difference between
the two is their focus: In contingency theory researchers are concerned with the optimal structuring of reporting relationships and decisionmaking responsibilities (e.g., Galbraith, 1973;
Lawrence & Lorsch, 1967), whereas in agency
theory they are concerned with the optimal
structuring of control relationships resulting
from these reporting and decision-making patterns. For example, using contingency theory,
we would be concerned with whether a firm is
organized in a divisional or matrix structure.
come-based contracts more attractive. In formal
terms,
Proposition 10: The length of the agency relationship is positively related to behavior-based
contracts and negatively related to outcomebased contracts.
Agency Theory and the
Organizational Literature
Despite Perrow’s (1986) assertion that agency
theory is very different from organization theory,
agency theory has several links to mainstream
organization perspectives (see Table 2). At its
roots, agency theory is consistent with the classic works of Barnard (1938) on the nature of cooperative behavior and March and Simon (1958)
on the inducements and contributions of the employment relationship. As in this earlier work,
the heart of agency theory is the goal conflict
inherent when individuals with differing preferences engage in cooperative effort, and the essential metaphor is that of the contract.
Agency theory is also similar to political models of organizations. Both agency and political
perspectives assume the pursuit of self-interest
at the individual level and goal conflict at the
organizational level (e.g., March, 1962; Pfeffer,
1981). Also, in both perspectives, information
Table 2
Comparison of Agency Theory Assumptions and Organizational Perspectives
Perspective
Assumption
Political
Self-interest
Goal conflict
Bounded rationality
Information asymmetry
Preeminence of efficiency
Risk aversion
Information as a commodity
X
X
Contingency
X
X
X
Organization
Control
Transaction
Cost
X
X
X
X
X
X
X
63
Agency
X
X
X
X
X
X
X
numbers bargaining. In agency theory there
are the risk attitudes of the principal and agent,
outcome uncertainty, and information systems.
Thus, the two theories share a parentage in economics, but each has its own focus and several
unique independent variables.
Using agency theory, we would be concerned
with whether managers within the chosen structure are compensated by performance incentives.
The most obvious tie is with the organizational
control literature (e.g., Dornbusch & Scott, 1974).
For example, Thompson’s (1967) and later Ouchi’s (1979) linking of known means/ends relationships and crystallized goals to behavior versus outcome control is very similar to agency
theory’s linking task programmability and measurability of outcomes to contract form (Eisenhardt, 1985). That is, known means/ends relationships (task programmability) lead to behavior control, and crystallized goals (measurable
outcomes) lead to outcome control. Similarly,
Ouchi’s (1979) extension of Thompson’s (1967)
framework to include clan control is similar to
assuming low goal conflict (Proposition 7) in
agency theory. Clan control implies goal congruence between people and, therefore, the reduced need to monitor behavior or outcomes.
Motivation issues disappear. The major differences between agency theory and the organizational control literature are the risk implications of principal and agent risk aversion and
outcome uncertainty (Propositions 4, 5, 6).
Not surprisingly, agency theory has similarities with the transaction cost perspective
(Williamson, 1975). As noted by Barney and Ouchi (1986), the theories share assumptions of selfinterest and bounded rationality. They also
have similar dependent variables; that is, hierarchies roughly correspond to behavior-based
contracts, and markets correspond to outcomebased contracts. However, the two theories
arise from different traditions in economics
(Spence, 1975): In transaction cost theorizing we
are concerned with organizational boundaries,
whereas in agency theorizing the contract between cooperating parties, regardless of boundary, is highlighted. However, the most important difference is that each theory includes
unique independent variables. In transaction
cost theory these are asset specificity and small
Contributions of Agency Theory
Agency theory reestablishes the importance
of incentives and self-interest in organizational
thinking (Perrow, 1986). Agency theory reminds
us that much of organizational life, whether we
like it or not, is based on self-interest. Agency
theory also emphasizes the importance of a
common problem structure across research topics. As Barney and Ouchi (1986) described it,
organization research has become increasingly
topic, rather than theory, centered. Agency theory reminds us that common problem structures
do exist across research domains. Therefore, results from one research area (e.g., vertical integration) may be germane to others with a common problem structure (e.g., compensation).
Agency theory also makes two specific contributions to organizational thinking. The first is the
treatment of information. In agency theory, information is regarded as a commodity: It has a
cost, and it can be purchased. This gives an
important role to formal information systems,
such as budgeting, MBO, and boards of directors, and informal ones, such as managerial
supervision, which is unique in organizational
research. The implication is that organizations
can invest in information systems in order to
control agent opportunism.
An illustration of this is executive compensation. A number of authors in this literature have
expressed surprise at the lack of performancebased executive compensation (e.g., Pearce,
Stevenson, & Perry, 1985; Ungson & Steers,
1984). However, from an agency perspective, it
is not surprising since such compensation
should be contingent upon a variety of factors
including information systems. Specifically,
64
viewed in terms of risk/reward trade-offs, not just
in terms of inability to preplan. The implication
is that outcome uncertainty coupled with differences in willingness to accept risk should influence contracts between principal and agent.
Vertical integration provides an illustration.
For example, Walker and Weber (1984) found
that technological and demand uncertainty did
not affect the “make or buy” decision for components in a large automobile manufacturer (principal in this case). The authors were unable to
explain their results using a transaction cost
framework. However, their results are consistent
with agency thinking if the managers of the automobile firm are risk neutral (a reasonable assumption given the size of the automobile firm
relative to the importance of any single component). According to agency theory, we would
predict that such a risk-neutral principal is relatively uninfluenced by outcome uncertainty,
which was Walker and Weber’s result.
Conversely, according to agency theory, the
reverse prediction is true for a new venture. In
this case, the firm is small and new, and it has
limited resources available to it for weathering
uncertainty: The likelihood of failure looms
large. In this case, the managers of the venture
may be risk-averse principals. If so, according to
agency theory we would predict that such managers will be very sensitive to outcome uncertainty. In particular, the managers would be
more likely to choose the “buy” option, thereby
transferring risk to the supplying firm. Overall,
agency theory predicts that risk-neutral managers are likely to choose the “make” option (behavior-based contract), whereas risk-averse executives are likely to choose “buy” (outcomebased contract).
richer information systems control managerial
opportunism and, therefore, lead to less performance-contingent pay.
One particularly relevant information system
for monitoring executive behaviors is the board
of directors. From an agency perspective, boards
can be used as monitoring devices for shareholder interests (Fama & Jensen, 1983). When
boards provide richer information, compensation is less likely to be based on firm performance. Rather, because the behaviors of top executives are better known, compensation based
on knowledge of executive behaviors is more
likely. Executives would then be rewarded for
taking well-conceived
actions (e.g., high
risk/high potential R&D) whose outcomes may
be unsuccessful. Also, when boards provide
richer information, top executives are more
likely to engage in behaviors that are consistent
with stockholders’ interests. For example, from
an agency viewpoint, behaviors such as using
greenmail and golden parachutes, which tend
to benefit the manager more than the stockholders, are less likely when boards are better monitors of stockholders’ interests. Operationally,
the richness of board information can be measured in terms of characteristics such as frequency of board meetings, number of board
subcommittees, number of board members with
long tenure, number of board members with
managerial and industry experience, and number of board members representing specific
ownership groups.
A second contribution of agency theory is its
risk implications. Organizations are assumed to
have uncertain futures. The future may bring
prosperity, bankruptcy, or some intermediate
outcome, and that future is only partly controlled
by organization members. Environmental effects such as government regulation, emergence of new competitors, and technical innovation can affect outcomes. Agency theory extends organizational thinking by pushing the
ramifications of outcome uncertainty to their
implications for creating risk. Uncertainty is
Empirical Results
Researchers in several disciplines have undertaken empirical studies of agency theory.
These studies, mirroring the two streams of theoretical agency research, are in Table 3.
65
to
(1) –
-.-
;5 0
to
0
U)
>
0
-0
0
0
U)
U)
U)
0
U)
1–
to
C
U)
tj
ti
>
–
.0
41:11
1–
(1)
7
0
m
>
cn
0
0)
U2
-U)
>
o
>
(1)
>
21)
0
0
ui
U)
m
0
U) U2
a) -0
>
0
U)
u
U)
U)
–
93
0
c:d
0
U2
ui a)
-0
U)
0
-0
U)
U)
0
0
7
0
-0
U)
0
U)
z
z
-U)
U)
U)
to
tm
tj
64
…4
Ui
>
(1)
(1)
-o
(1)
U)
1-0
a)
0)
ti
U)
(1)
0
>
-,-
U)
r4
V
1–
o
04
(1) 0
04
J–
5
:
tD)
.-
–
(1)
U
0
U)
(1)
F4
0
0
U)
C:d
tD)
-w
(1)
0
(1)
U)
0
(1) U)
>
o
U)
U)
-4
–
—
U)
(1)
I-,
ti
cn 0 CD
U)
” CY)
U) U)
ri, CD
CF)
(1)
0
U)
U)
U)
Q)U)
r,
CY)
CD
LO
LO
CY)
0
U)
”
o
U)
U)
U) _C:
0 — 0
_C:
U)
LO
CD
ui
(1)(1)-),
0
LO
CD
U)
U)
.-
0
-C:
(1)
U)
0
C:d
CF)
CY)
0
0
cr)
u
CD
ti
U)
to
>
U)
0
U)
>
U)
0
00
to
v
CV)
U)
>
U)
0
00
CF)
00
CF)
>
U)
0
LO
00
CF)
_z:
0
U)
0
U)
(1)
U)
66
00
00
U)
(1)
U)
0
(1)
00
CF)
ui
a)
0
tj
‘O
ti
o
ui
a) U2
a)
U2
-.1
U2
U2
r-I
a) a)
0
o r-I
0
0
a)
0
0
‘U2
0
ui
C/)
0
-0
0
U2
ts 0
0
0
0
U2
U2
0 -Z4
0) 0
C:6-.1
a)
0)
U2
N
U2
C/)
a)
a)
>
ui
ui
U2
a)
t
-uJ
o o
w
U2
tj
“O
CD
0
0
0
ui
a)
a)
>,
0
U2
0
U2
U2
T:
r:
0
a)
U2
k4
(
U-2
o
a)
>
t))
U2
o
a) >
0 a)
a)
U2
_C:
CD
CD
a)
0
-.1
U2
0
.1
04
o
a)
U2
06
Cd (1)
0 -r–o –.z
0
1
r- ” 0 (1) &.
0
U2 .-u
U2
r.
a)
a)
a) r-,
0
a)
rU2
0 >1
U2
U2
a)
t))
“W
CD
Clq
CY)
00
a)
0
a) d
>
ti
co
PC
I.-
u
–U2
–U2
U2
U2
0
U2
0
U2
0
U2
0
-.1
U2
I-
>
>
00
0
0
rn
U2
—
>
cr)
F!
ti
U2
0
4
a) a)
00
00
iz
00
cr)
a)
00
0 cr)
v
U
m
_U2
w
67
U2
0
-.1
00
co
a)
:,:
U2
a)
U2
U2
J,-
Results of the Positivist Stream
their firms (outcome-based contracts) were less
likely to resist takeover bids.
The effects of market discipline on agency relationships were examined in Wolfson’s (1985)
study of the relationship between the limited
(principals) and general (agent) partners in oil
and gas tax shelter programs. In this study, both
tax and agency effects were combined in order
to assess why the limited partnership governance form survived in this setting despite extensive information advantages and divergent
incentives for the limited partner. Consistent
with agency arguments (Fama, 1980), Wolfson
found that long-run reputation effects of the market coaligned the short-run behaviors of the general partner with the limited partners’ welfare.
Kosnik (1987) examined another information
mechanism for managerial opportunism, the
board of directors. Kosnik studied 110 large U. S.
corporations that were greenmail targets between 1979 and 1983. Using both hegemony and
agency theories, she related board characteristics to whether greenmail was actually paid
(paying greenmail is considered not in the stockholders’ interests). As predicted by agency theory (Fama & Jensen, 1983), boards of companies
that resisted greenmail had a higher proportion
of outside directors and a higher proportion of
outside director executives.
In a similar vein, Argawal and Mandelker
(1987) examined whether executive holdings of
firm securities reduced agency problems between stockholders and management. Specifically, they studied the relationship between
stock and stock option holdings of executives
and whether acquisition and financing decisions were made consistent with the interests of
stockholders. In general, managers prefer lower
risk acquisitions and lower debt financing (see
Argawal & Mandelker, 1987, for a review). Their
sample included 209 firms that participated in
acquisitions and divestitures between 1974 and
1982. Consistent with agency ideas (e.g., Jensen
& Meckling, 1976), executive security holdings
(outcome-based contract) were related to acqui-
In the positivist stream, the common approach
is to identify a policy or behavior in which stockholder and management interests diverge and
then to demonstrate that information systems or
incentives solve the agency
outcome-based
problem. That is, these mechanisms coalign
managerial behaviors with owner preferences.
Consistent with the positivist tradition, most of
these studies concern the separation of ownership from management in large corporations,
and they use secondary source data that are
available for large firms.
One of the earliest studies of this type was
conducted by Amihud and Lev (1981). These researchers explored why firms engage in conglomerate mergers. In general, conglomerate
mergers are not in the interests of the stockholders because, typically, stockholders can diversify directly through their stock portfolio. In contrast, conglomerate mergers may be attractive
to managers who have fewer avenues available
to diversify their own risk. Hence, conglomerate
mergers are an arena in which owner and manager interests diverge. Specifically, these authors linked merger and diversification behaviors to whether the firm was owner controlled (i.e.,
had a major stockholder) or manager controlled
(i.e., had no major stockholder). Consistent with
agency theory arguments (Jensen & Meckling,
1976), manager-controlled firms engaged in significantly more conglomerate (but not more related) acquisitions and were more diversified.
Along the same lines, Walking and Long
(1984) studied managers’ resistance to takeover
bids. Their sample included 105 large U.S. corporations that were targets of takeover attempts
between 1972 and 1977. In general, resistance to
takeover bids is not in the stockholders’ interests,
but it may be in the interests of managers because they can lose their jobs during a takeover.
Consistent with agency theory (Jensen & Meckling, 1976), the authors found that managers
who have substantial equity positions within
68
port the positivist propositions described earlier.
Similarly, laboratory studies by Dejong and colleagues (1985), which are not reviewed here,
are also supportive.
sition and financing decisions that were more
consistent with stockholder interest. That is, executive stock holdings appeared to coalign
managerial preferences with those of stockholders.
Singh and Harianto (in press) studied golden
parachutes in a matched sample of 84 Fortune
500 firms. Their study included variables from
both agency and managerialist perspectives.
Consistent with agency theory (Jensen & Meckling, 1976; Fama & Jensen, 1983), the authors
found that golden parachutes are used to coalign
executive interests with those of stockholders in
takeover situations, and they are seen as an alternative outcome-based contract to executive
stock ownership. Specifically, the authors found
that golden parachutes were positively associated with a higher probability of a takeover attempt and negatively associated with executive
stock holdings.
Finally, Barney (1988) explored whether employee stock ownership reduces a firm’s cost of
equity capital. Consistent with agency theory
(Jensen & Meckling, 1976), Barney argued that
employee stock ownership (outcome-based contract) would coalign the interests of employees
with stockholders. Using efficient capital market
assumptions, he further argued that this coalignment would be reflected in the market through a
lower cost of equity. Although Barney did not
directly test the agency argument, the results
are consistent with an agency view.
In summary, there is support for the existence
of agency problems between shareholders and
top executives across situations in which their
interests diverge-that
is, takeover attempts,
debt versus equity financing, acquisitions, and
divestitures, and for the mitigation of agency
problems (a) through outcome-based contracts
such as golden parachutes (Singh & Harianto,
in press) and executive stock holdings (Argawal
& Mandelker, 1987; Walking & Long, 1984) and
(b) through information systems such as boards
(Kosnik, 1987) and efficient markets (Barney,
1988; Wolfson, 1985). Overall, these studies sup-
Results of the Principal-Agent Stream
The principal-agent stream is more directly focused on the contract between the principal and
the agent. Whereas the positivist stream lays the
foundation (that is, that agency problems exist
and that various contract alternatives are available), the principal-agent stream indicates the
most efficient contract alternative in a given situation. The common approach in these studies
is to use a subset of agency variables such as
task programmability, information systems, and
outcome uncertainty to predict whether the contract is behavior- or outcome-based. The underlying assumption is that principals and agents
will choose the most efficient contract, although
efficiency is not directly tested.
In one study, Anderson (1985) probed vertical
integration using a transaction cost perspective
with agency variables. Specifically, she examined the choice between a manufacturer’s representative (outcome-based) and a corporate
sales force (behavior-based) among a sample of
electronics firms. The most powerful explanatory variable was from agency theory: the difficulty of measuring outcomes (measured by
amount of nonselling tasks and joint team sales).
Consistent with agency predictions, this variable was positively related to using a corporate
sales force (behavior-based contract).
In other studies, Eisenhardt (1985, 1988) examined the choice between commission (outcomebased) and salary (behavior-based) compensation of salespeople in retailing. The original
study (1985) included only agency variables,
while a later study (1988) added additional
agency variables and institutional theory predictions. The results supported agency theory
predictions that task programmability, information systems (measured by the span of control),
and outcome uncertainty variables (measured
69
has empirical support. Overall, it seems reasonable to urge the adoption of an agency theory
perspective when investigating the many problems that have a principal-agent structure. Five
specific recommendations are outlined below
for using agency theory in organizational research.
by number of competitors and failure rates) significantly predict the salary versus commission
choice. Institutional variables were significant
as well.
Conlon and Parks (1988) replicated and extended Eisenhardt’s work in a laboratory setting. They used a multiperiod design to test both
agency and institutional predictions. Consistent
with agency theory (Harris & Raviv, 1978), they
found that information systems (manipulated by
whether or not the principal could monitor the
agent’s behavior) were negatively related to
performance-contingent (outcome-based) pay.
They also found support for the institutional predictions.
Finally, Eccles (1985) used agency theory to
develop a framework for understanding transfer
pricing. Using interviews with 150 executives in
13 large corporations, he developed a framework based on notions of agency and fairness to
prescribe the conditions under which various
sourcing and transfer pricing alternatives are
both efficient and equitable. Prominent in his
framework is the link between decentralization
(arguably a measure of task programmability)
and the choice between cost (behavior-based
contract) and market (outcome-based contract)
transfer pricing mechanisms.
In summary, there is support for the principalagent hypotheses linking contract form with (a)
information systems (Conlon & Parks, 1988; Eccles, 1985; Eisenhardt, 1985), (b) outcome uncertainty (Eisenhardt, 1985), (c) outcome measurability (Anderson, 1985; Eisenhardt, 1985), (d)
time (Conlon & Parks, 1988), and (e) task programmability (Eccles, 1985; Eisenhardt, 1985).
Moreover, this support rests on research using a
variety of methods including questionnaires,
secondary sources, laboratory experiments,
and interviews.
Focus on Information Systems, Outcome
Uncertainty, and Risk
McGrath, Martin, and Kukla (1981) argued
that research is a knowledge accrual process.
Using this accrual criterion, next steps for
agency theory research are clear: Researchers
should focus on information systems, outcome
uncertainty, and risk. These agency variables
make the most unique contribution to organizational research, yet they have received little empirical attention (Table 3). It is important that researchers place emphasis on these variables in
order to advance agency theory and to provide
new concepts in the study of familiar topics such
as impression management, innovation, vertical integration, compensation, strategic alliances, and board relationships.
Studying risk and outcome uncertainty is particularly opportune because of recent advances
in measuring risk preferences. By relying on the
works of Kahneman and Tversky (1979), MacCrimmon and Wehrung (1986), and March and
Shapira (1987), the organizational researcher
can measure risk preference more easily and
realistically. These techniques include direct
measures of risk preference such as lotteries and
indirect measures using demographic characteristics such as age and wealth and payoff
characteristics such as gain versus loss. (See
March and Shapira, 1987, for a review.)
Key on Theory-Relevant Contexts
Organizational theory usually is explored in
settings in which the theory appears to have
greatest relevance. For example, institutional
and resource dependence theories were developed primarily in large, public bureaucracies in
which efficiency may not have been a pressing
Recommendations for Agency
Theory Research
As argued above, agency theory makes contributions to organization theory, is testable, and
70
The second area is expansion beyond the
pure forms of behavior and outcome contracts as
described in this article to a broader range of contract altematives. Most research (e.g., Anderson,
1985; Eisenhardt, 1985, 1988) treats contracts as a
dichotomy: behavior versus outcome. However,
contracts can vary on a continuum between behavior and outcome contracts. Also, current research focuses on a single reward, neglecting
many situations in which there are multiple rewards, differing by time frame and contract basis. For example, upper level managers usually
are compensated through multiple rewards
such as promotions, stock options, and salary.
Both multiple and mixed rewards (behavior and
outcome) present empirical difficulties, but they
also mirror real life. The richness and complexity of agency theory would be enhanced if researchers would consider this broader spectrum
of possible contracts.
concern. The recommendation here is to take
the same approach with agency theory: Key on
theory-relevant contexts.
Agency theory is most relevant in situations in
which contracting problems are difficult. These
include situations in which there is (a) substantial goal conflict between principals and agents,
such that agent opportunism is likely (e.g., owners and managers, managers and professionals, suppliers and buyers); (b) sufficient outcome
uncertainty to trigger the risk implications of the
theory (e.g., new product innovation, young
and small firms, recently deregulated industries); and (c) unprogrammed or team-oriented
jobs in which evaluation of behaviors is difficult.
By emphasizing these contexts, researchers can
use agency theory where it can provide the most
leverage and where it can be most rigorously
tested. Topics such as innovation and settings
such as technology-based firms are particularly
attractive because they combine goal conflict
between professionals and managers, risk, and
jobs in which performance evaluation is difficult.
Use Multiple Theories
A recent article by Hirsch et al. (1987) eloquently compared economics with sociology.
They argued that economics is dominated by a
single paradigm, price theory, and a single
view of human nature, self-interest. In contrast,
the authors maintained that a strength of organizational research is its polyglot of theories that
yields a more realistic view of organizations.
Consistent with the Hirsch et al. arguments,
the recommendation here is to use agency theory with complementary theories. Agency theory presents a partial view of the world that,
although it is valid, also ignores a good bit of the
complexity of organizations. Additional perspectives can help to capture the greater complexity.
This point is demonstrated by many of the empirical studies reviewed above. For example,
the Singh and Harianto (in press) and Kosnik
(1987) studies support agency theory hypotheses, but they also use the complementary perspectives of hegemony and managerialism.
These perspectives emphasize the power and political aspects of golden parachutes and green-
Expand to Richer Contexts
Perrow (1986)and others have criticized agency
theory for being excessively narrow and having
few testable implications. Although these criticisms may be extreme, they do suggest that research should be undertaken in new areas.
Thus, the recommendation is to expand to a
richer and more complex range of contexts.
Two areas are particularly appropriate. One
is to apply the agency structure to organizational behavior topics that relate to information
asymmetry (or deception) in cooperative situations. Examples of such topics are impression
management (Gardner & Martinko, 1988), lying
and other forms of secrecy (Sitkin, 1987), and
blame (Leatherwood & Conlon, 1987). Agency
theory might contribute an overall framework in
which to place these various forms of selfinterest, leading to a better understanding of
when such behaviors will be likely and when
they will be effective.
71
mail, respectively. Similarly, the studies by
Eisenhardt (1988) and Conlon and Parks (1988)
combine institutional and agency theories. The
institutional emphasis on tradition complements
the efficiency emphasis of agency theory, and
the result is a better understanding of compensation. Other examples include Anderson (1985),
who coupled agency and transaction cost, and
Eccles (1985), who combined agency with equity
theory.
style is to risk doing second-rate economics without contributing first-rate organizational research. Therefore, although it is appropriate to
monitor developments in economics, it is more
useful to treat economics as an adjunct to more
mainstream empirical work by organizational
scholars.
Look Beyond Economics
This paper began with two extreme positions
on agency theory-one
arguing that agency
theory is revolutionary and a powerful foundation (Jensen, 1983) and the other arguing that the
theory addresses no clear problem, is narrow,
lacks testable implications, and is dangerous
(Perrow, 1986). A more valid perspective lies in
the middle. Agency theory provides a unique,
realistic, and empirically testable perspective
on problems of cooperative effort. The intent of
this paper is to clarify some of the confusion surrounding agency theory and to lead organizational scholars to use agency theory in their
study of the broad range of principal-agent issues facing firms.
Conclusion
The final recommendation is that organizational researchers should look beyond the economics literature. The advantages of economics
are careful development of assumptions and
logical propositions (Hirsch et al., 1987). However, much of this careful theoretical development has already been accomplished for agency
theory. For organizational researchers, the payoff now is in empirical research, where organizational researchers have comparative advantage (Hirsch et al., 1987). To rely too heavily on
economics with its restrictive assumptions such
as efficient markets and its single-perspective
References
Anderson, E. (1985) The salesperson as outside agent of employee: A transaction cost analysis. Marketing Science, 4,
234-254.
Basu, A., Lal, R., Srinivasan, V., & Staelin, R. (1985) Salesforce compensation plans: An agency theoretic perspective. Marketing Science, 4, 267-291.
Amihud, Y., & Lev, B. (1981) Risk reduction as a managerial
motive for conglomerate mergers. Bell Journal of Economics, 12, 605-616.
Berle, A., & Means, G. (1932) The modern corporation and
private property. New York: Macmillan.
Bolton, M. (1988) Organizational miming: When do late
adopters of organizational innovations outperform pioneers? Paper presented at the meeting of the Academy of
Management, Anaheim, CA.
Argawal, A., & Mandelker, G. (1987) Managerial incentives
and corporate investment and financing decisions. Journal of Finance, 42, 823-837.
Arrow, K. (1971) Essays in the theory of risk bearing. Chicago: Markham.
Barnard, C. (1938) The functions of the executive. Cambridge, MA: Harvard University Press.
Barney, J. (1988) Agency theory, employee stock ownership
and a firm’s cost of equity capital. Unpublished working
paper, Texas A&M University, College Station.
Barney, J., & Ouchi, W. (Eds.) (1986) Organizational economics. San Francisco: Jossey-Bass.
Burt, R. (1979) A structural theory of interlocking corporate
directorates. Social Networks, 1, 415-435.
Capitalism in the making. (1984, April 30) Time, p. 62.
Chandler, A. (1962) Strategy and structure. New York: Doubleday.
Conlon, E., & Parks, J. (1988) The effects of monitoring and
tradition on compensation arrangements: An experiment
on principal/agent dyads. In F. Hoy (Ed.), Best papers pro-
72
ceedings (pp. 191-195). Anaheim, CA: Academy of Management.
Holmstrom, B. (1979) Moral hazard and observability. Bell
Journal of Economics, 10, 74-91.
Cyert, R., & March, J. (1963) A behavioral theory of the firm.
Englewood Cliffs, NJ: Prentice-Hall.
Jensen, M. (1983) Organization theory and methodology. Accounting Review, 56, 319-338.
Dejong, D., Forsythe, R., & Uecker, W. (1985) Ripoffs, lemons
and reputation formation in agency relationships: A laboratory market study. Journal of Finance, 50, 809-820.
Jensen, M. (1984) Takeovers: Folklore and science. Harvard
Business Review, 62(6), 109-121.
Jensen, M., & Meckling, W. (1976) Theory of the firm: Managerial behavior, agency costs, and ownership structure.
Journal of Financial Economics, 3, 305-360.
Demski, J. (1980) A simple case of indeterminate financial
reporting. Working paper, Stanford University.
Jensen, M., & Roeback, R. (1983) The market for corporate
control: Empirical evidence. Journal of Financial Economics, 11, 5-50.
Demski, J., & Feltham, G. (1978) Economic incentives in budgetary control systems. Accounting Review, 53, 336-359.
Dornbusch, S., & Scott, W. R. (1974) Evaluation and the exercise of authority. San Francisco: Jossey-Bass.
Kahneman, D., & Tversky, A. (1979) Prospect theory: An
analysis of decisions under risk. Econometrica, 47, 263291.
Eccles, R. (1985) Transfer pricing as a problem of agency. In
J. Pratt & R. Zeckhauser (Eds.), Principals and agents: The
structure of business (pp. 151-186). Boston: Harvard Business School Press.
Kosnik, R. (1987) Greenmail: A study in board performance
in corporate governance. Administrative Science Quarterly, 32, 163-185.
Eisenhardt, K. (1985) Control: Organizational and economic
approaches. Management Science, 31, 134- 149.
Lambert, R. (1983) Long-term contracts and moral hazard.
Bell Journal of Economics, 14, 441-452.
Eisenhardt, K. (1988) Agency and institutional explanations
of compensation in retail sales. Academy of Management
Journal, 31, 488-511.
Lawrence, P., & Lorsch, J. (1967) Organization and environment. Boston: Division of Research, Harvard Business
School.
Fama, E. (1980) Agency problems and the theory of the firm.
Journal of Political Economy, 88, 288-307.
Fama, E., & Jensen, M. (1983) Separation of ownership and
control. Journal of Law and Economics, 26, 301-325.
Leatherwood, M., & Conlon, E. (1987) Diffusibility of blame:
Effects on persistence in a project. Academy of Management Journal, 30, 836-848.
Galbraith, J. (1973) Designing complex organizations. Reading, MA: Addison-Wesley.
MacCrimmon, K., & Wehrung, D. (1986) Taking risks: The
management of uncertainty. New York: Free Press.
Gardner, W., & Martinko, M. (1988) Impression management: An observational study linking audience characteristics with verbal self-presentations. Academy of Management Journal, 31, 42-65.
March, J. (1962) The business firm as a political coalition.
Journal of Politics, 24, 662-678.
Gausch, J., & Weiss, A. (1981) Self-selection in the labor market. American Economic Review, 71, 275-284.
March, J., & Simon, H. (1958) Organizations.
Wiley.
Harris, M., & Raviv, A. (1978) Some results on incentive contracts with application to education and employment,
health insurance, and law enforcement. American Economic Review, 68, 20-30.
McGrath, J., Martin, J., & Kukla, R. (1982) Judgment calls in
research. Beverly Hills, CA: Sage.
Harris, M., & Raviv, A. (1979) Optimal incentive contracts
with imperfect information. Journal of Economic Theory,
20, 231-259.
Mitnick, B. (1986) The theory of agency and organizational
analysis. Unpublished working paper, University of Pittsburgh.
Hirsch, P., & Friedman, R. (1986) Collaboration or paradigm
shift? Economic vs. behavioral thinking about policy? In J.
Pearce & R. Robinson (Eds.), Best papers proceedings (pp.
31-35). Chicago: Academy of Management.
Ouchi, W. (1979) A conceptual framework for the design of
organizational control mechanisms. Management Science, 25, 833-848.
March, J., & Shapira, Z. (1987) Managerial perspectives on
risk and risk taking. Management Science, 33, 1404-1418.
New York:
Meet Mike Jensen, the professor of merger mania. (1988,
February 8) Business Week, pp. 66-67.
Pearce, J., Stevenson, W., & Perry, J. (1985) Managerial compensation based on organizational performance: A time
series analysis of the effects of merit pay. Academy of
Management journal, 28, 261-278.
Hirsch, P., Michaels, S., & Friedman, R. (1987) “Dirty hands”
versus “clean models”: Is sociology in danger of being
seduced by economics? Theory and Society, 317-336.
73
Ungson, G., & Steers, R. (1984) Motivation and politics in
executive compensation. Academy of Management Review, 9, 313-323.
Perrow, C. (1986) Complex organizations. New York: Random House.
Pettigrew, A. (1973) The politics of organizational decision
making. London: Tavistock.
Walker, G., & Weber, D. (1984) A transaction cost approach
to make-or-buy decisions. Administrative Science Quarterly, 29, 373-391.
Pfeffer, J. (1981) Power in organizations. Marshfield, MA: Pittman.
Pfeffer, J., & Salancik, G. (1974) Organizational decision
making as a political process: The case of a university
budget. Administrative Science Quarterly, 19, 135-151.
Walking, R., & Long, M. (1984) Agency theory, managerial
welfare, and takeover bid resistance. The Rand Journal of
Economics, 15, 54-68.
Ross, S. (1973) The economic theory of agency: The principal’s problem. American Economic Review, 63, 134-139.
White, H. (1985) Agency as control. In J. Pratt & R. Zeckhauser (Eds.), Principals and agents: The structure of business (pp. 187-214). Boston: Harvard Business School
Press.
Shavell, S. (1979) Risk sharing and incentives in the principal
and agent relationship. Bell Journal of Economics, 10, 5373.
Singh, H., & Harianto, F. (in press) Management-board relationships, takeover risk and the adoption of golden parachutes: An empirical investigation. Academy of Management Journal.
Williamson, 0. (1975) Markets and hierarchies: Analysis and
antitrust implications. New York: Free Press.
Wilson, R. (1968) On the theory of syndicates. Econometrica,
36, 119-132.
Sitkin, S. (1987) Secrecy in organizations: The limits of legitimate information control. Working paper, University of
Texas, Austin.
Wolfson, M. (1985) Empirical evidence of incentive problems
and their mitigation in oil and gas shelter programs. In J.
Pratt & R. Zeckhauser (Eds.), Principals and agents: The
structure of business (pp. 101- 126). Boston: Harvard Business School Press.
Spence, A. M. (1975) The economics of internal organization:
An introduction. Bell Journal of Economics, 6, 163-172.
Spence, A. M., & Zeckhauser, R. (1971) Insurance, information, and individual action. American Economic Review,
61, 380-387.
Zenger, T. (1988)Agency sorting, agent solutions and diseconomies of scale: An empirical investigation of employment
contracts in high technology R&D. Paper presented at the
meeting of the Academy of Management, Anaheim, CA.
Thompson, J. (1967) Organizations in action. New York: McGraw-Hill.
Kathleen M. Eisenhardt (Ph.D., Stanford University) is
Assistant Professor at Stanford University. Correspondence can be sent to her at the Department of Industrial Engineering and Engineering Management, 346
Terman Building, Stanford University, Stanford, CA
94305.
The author thanks Paul Adler, Michele Bolton, Philip
Bromiley, Jim Hodder, William Ouchi, Gerald Salancik, Kaye Schoonhoven, and Robert Sutton for their
comments and suggestions.
74
Prospect Theory: An Analysis of Decision under Risk
Daniel Kahneman; Amos Tversky
Econometrica, Vol. 47, No. 2. (Mar., 1979), pp. 263-292.
Stable URL:
http://links.jstor.org/sici?sici=0012-9682%28197903%2947%3A2%3C263%3APTAAOD%3E2.0.CO%3B2-3
Econometrica is currently published by The Econometric Society.
Your use of the JSTOR archive indicates your acceptance of JSTOR’s Terms and Conditions of Use, available at
http://www.jstor.org/about/terms.html. JSTOR’s Terms and Conditions of Use provides, in part, that unless you have obtained
prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in
the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/journals/econosoc.html.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is an independent not-for-profit organization dedicated to and preserving a digital archive of scholarly journals. For
more information regarding JSTOR, please contact support@jstor.org.
http://www.jstor.org
Thu Apr 5 17:18:55 2007
ECONOMETRICA
PROSPECT THEORY: AN ANALYSIS O F DECISION UNDER RISK
This paper presents a critique of expected utility theory as a descriptive model of
decision making under risk, and develops an alternative model, called prospect theory.
Choices among risky prospects exhibit several pervasive effects that are inconsistent with
the basic tenets of utility theory. In particular, people underweight outcomes that are
merely probable in comparison with outcomes that are obtained with certainty. This
tendency, called the certainty effect, contributes to risk aversion in choices involving sure
gains and to risk seeking in choices involving sure losses. In addition, people generally
discard components that are shared by all prospects under consideration. This tendency,
called the isolation effect, leads to inconsistent preferences when the same choice is
presented in different forms. An alternative theory of choice is developed, in which value
is assigned to gains and losses rather than to final assets and in which probabilities are
replaced by decision weights. The value function is normally concave for gains, commonly
convex for losses, and is generally steeper for losses than for gains. Decision weights are
generally lower than the corresponding probabilities, except in the range of low probabilities. Overweighting of low probabilities may contribute to the attractiveness of both
insurance and gambling.
1. INTRODUCTION
EXPECTEDUTILITY THEORY has dominated the analysis of decision making under
risk. It has been generally accepted as a normative model of rational choice [24],
and widely applied as a descriptive model of economic behavior, e.g. [15,4].
Thus, it is assumed that all reasonable people would wish to obey the axioms of the
theory [47,36], and that most people actually do, most of the time.
The present paper describes several classes of choice problems in which
preferences systematically violate the axioms of expected utility theory. In the
light of these observations we argue that utility theory, as it is commonly
interpreted and applied, is not an adequate descriptive model and we propose an
alternative account of choice under risk.
2. CRITIQUE
Decision making under risk can be viewed as a choice between prospects or
gambles. A prospect (xl, pl; . .. ;x,,, p,) is a contract that yields outcome xi with
probability pi, where pl + p z + . . . +p, = 1. To simplify notation, we omit null
outcomes and use (x, p) to denote the prospect (x, p ; 0, 1 -p) that yields x with
probability p and 0 with probability 1-p. The (riskless) prospect that yields x
with certainty is denoted by (x). The present discussion is restricted to prospects
with so-called objective or standard probabilities.
The application of expected utility theory to choices between prospects is based
on the following three tenets.
(i) Expectation: U(x1, p ~. ;. . ;x,, p,) = p l u ( x l ) + . . . +p,u(x,).
‘
This work was supported in part by grants from the Harry F. Guggenheim Foundation and from
the Advanced Research Projects Agency of the Department of Defense and was monitored by Office
of Naval Research under Contract N00014-78-C-0100 (ARPA Order No. 3469) under Subcontract
78-072-0722 from Decisions and Designs, Inc. to Perceptronics, Inc. We also thank the Center for
Advanced Study in the Behavioral Sciences at Stanford for its support.
263
264
D. KAHNEMAN AND A. TVERSKY
That is, the overall utility of a prospect, denoted by U, is the expected utility of
its outcomes.
(ii) Asset Integration: (xl, pl; . . . ;x,, p,) is acceptable at asset position w iff
U(w + X I ,p1;. . . ; w +x”, p,)> u(w).
That is, a prospect is acceptable if the utility resulting from integrating the
prospect with one’s assets exceeds the utility of those assets alone. Thus, the
domain of the utility function is final states (which include one’s asset position)
rather than gains or losses.
Although the domain of the utility function is not limited to any particular class
of consequences, most applications of the theory have been concerned with
monetary outcomes. Furthermore, most economic applications introduce the
following additional assumption.
(iii) Risk Aversion: u is concave (u”< 0). A person is risk averse if he prefers the certain prospect (x) to any risky prospect with expected value x. In expected utility theory, risk aversion is equivalent to the concavity of the utility function. The prevalence'of risk aversion is perhaps the best known generalization regarding risky choices. It led the early decision theorists of the eighteenth century to propose that utility is a concave function of money, and this idea has been retained in modern treatments (Pratt [33], Arrow [41). In the following sections we demonstrate several phenomena which violate these tenets of expected utility theory. The demonstrations are based on the responses of students and university faculty to hypothetical choice problems. The respondents were presented with problems of the type illustrated below. Which of the following would you prefer? A: 50°/o chance to win 1,000, B: 450 for sure. 50% chance to win nothing; The outcomes refer to Israeli currency. To appreciate the significance of the amounts involved, note that the median net monthly income for a family is about 3,000 Israeli pounds. The respondents were asked to imagine that they were actually faced with the choice described in the problem, and to indicate the decision they would have made in such a case. The responses were anonymous, and the instructions specified that there was no 'correct' answer to such problems, and that the aim of the study was to find out how people choose among risky prospects. The problems were presented in questionnaire form, with at most a dozen problems per booklet. Several forms of each questionnaire were constructed so that subjects were exposed to the problems in different orders. In addition, two versions of each problem were used in which the left-right position of the prospects was reversed. The problems described in this paper are selected illustrations of a series of effects. Every effect has been observed in several problems with different outcomes and probabilities. Some of the problems have also been presented to groups of students and faculty at the University of Stockholm and at the 265 PROSPECT THEORY University of Michigan. The pattern of results was essentially identical to the results obtained from Israeli subjects. The reliance on hypothetical choices raises obvious questions regarding the validity of the method and the generalizability of the results. We are keenly aware of these problems. However, all other methods that have been used to test utility theory also suffer from severe drawbacks. Real choices can be investigated either in the field, by naturalistic or statistical observations of economic behavior, or in the laboratory. Field studies can only provide for rather crude tests of qualitative predictions, because probabilities and utilities cannot be adequately measured in such contexts. Laboratory experiments have been designed to obtain precise measures of utility and probability from actual choices, but these experimental studies typically involve contrived gambles for small stakes, and a large number of repetitions of very similar problems. These features of laboratory gambling complicate the interpretation of the results and restrict their generality. By default, the method of hypothetical choices emerges as the simplest procedure by which a large number of theoretical questions can be investigated. The use of the method relies on the assumption that people often know how they would behave in actual situations of choice, and on the further assumption that the subjects have no special reason to disguise their true preferences. If people are reasonably accurate in predicting their choices, the presence of common and systematic violations of expected utility theory in hypothetical problems provides presumptive evidence against that theory. Certainty, Probability, and Possibility In expected utility theory, the utilities of outcomes are weighted by their probabilities. The present section describes a series of choice problems in which people's preferences systematically violate this principle. We first show that people overweight outcomes that are considered certain, relative to outcomes which are merely probable-a phenomenon which we label the certainty effect. The best known counter-example to expected utility theory which ekploits the certainty effect was introduced by the French economist Maurice Allais in 1953 [2]. Allais' example has been discussed from both normative and descriptive standpoints by many authors [28,38].The following pair of choice problems is a variation of Allais' example, which differs from the original in that it refers to moderate rather than to extremely large gains. The number of respondents who answered each problem is denoted by N, and the percentage who choose each option is given in brackets. PROBLEM 1: Choose between A: 2,500 with probability .33, 2,400 with probability .66, 0 with probability .01; N=72 [I81 B: 2,400 with certainty. 266 D. KAMNEMAN AND A. TVERSKY PROBLEM2: Choose between C: 2,500 with probability .33, 0 with probability .67; D: 2,400 with probability .34, 0 with probability .66. The data show that 82 per cent of the subjects chose B in Problem 1, and 83 per cent of the subjects chose C in Problem 2. Each of these preferences is significant at the .O1 level, as denoted by the asterisk. Moreover, the analysis of individual patterns of choice indicates that a majority of respondents (61 per cent) made the modal choice in both problems. This pattern of preferences violates expected utility theory in the manner originally described by Allais. According to that theory, with u(0) = 0, the first preference implies while the second preference implies the reverse inequality. Note that Problem 2 is obtained from Problem 1by eliminating a .66 chance of winning 2400 from both prospects. under consideration. Evidently, this change produces a greater reduction in desirability when it alters the character of the prospect from a sure gain to a probable one, than when both the original and the reduced prospects are uncertain. A simpler demonstration of the same phenomenon, involving only twooutcome gambles is given below. This example is also based on Allais [2]. In this pair of problems as well as in all other problem-pairs in this section, over half the respondents violated expected utility theory. To show that the modal pattern of preferences in Problems 3 and 4 is not compatible with the theory, set u(0) = 0, and recall that the choice of B implies u(3,000)/u(4,000)>4/5,
whereas the choice of C implies the reverse inequality. Note that the prospect
C = (4,000, .20) can be expressed as (A, .25), while the prospect D = (3,000, .25)
can be rewritten as (B,.25). The substitution axiom of utility theory asserts that if
B is preferred to A, then any (probability) mixture (B, p) must be preferred to the
mixture (A, p). Our subjects did not obey this axiom. Apparently, reducing the
probability of winning from 1.0 to .25 has a greater effect than the reduction from
267
PROSPECT THEORY
.8 to .2. The following pair of choice problems illustrates the certainty effect with
non-monetary outcomes.
A:
50% chance to win a threeweek tour of England,
France, and Italy;
C: 5% chance to win a threeweek tour of England,
France, and Italy;
N=72
[671*
B:
A one-week tour of
England, with certainty.
D:
10% chance to win a oneweek tour of England.
[331
The certainty effect is not the only type of violation of the substitution axiom.
Another situation in which this axiom fails is illustrated by the followingproblems.
Note that in Problem 7 the probabilities of winning are substantial (.90 and .45),
and most people choose the prospect where winning is more probable. In Problem
8, there is a possibility of winning, although the probabilities of winning are
minuscule (.002 and .001) in both prospects. In this situation where winning is
possible but not probable, most people choose the prospect that offers the larger
gain. Similar results have been reported by MacCrimmon and Larsson [28].
The above problems illustrate common attitudes toward risk or chance that
cannot be captured by the expected utility model. The results suggest the
following empirical generalization concerning the manner in which the substitution axiom is violated. If ( y , pq) is equivalent to (x, p), then ( y , pqr) is preferred to
(x, pr), 0 to denote the
prevalent preference, i.e., the choice made by the majority of subjects.
TABLE I
Posltlve prospects
Problem 3:
N=95
Problem 4 :
N = 95
Problem 7 :
N=66
Problem 8:
N=66
(4,000, .80) < (3,000). [201 (4,000, .20) > (3,000, .25).
[651*
[351
(3,000, .90) > (6,000, .45).
[86]*
[I41
(3,000, ,002) < (6,000, ,001). ~271 [731* Negative prospects Problem 3': (-4,000, .80) 1 (-3,000). N=95 [921* [81 Problem 4': (-4,000, .20) < (-3,000, .25). N=95 [421 [581 Problem 7 ' : (-3,000, .90) < (-6,000, .45). N=66 [81 r921* Problem 8': (-3,000, ,002) > (-6,000, ,001).
N=66
r701*
~301
In each of the four problems in Table I the preference between negative
prospects is the mirror image of the preference between positive prospects. Thus,
the reflection of prospects around 0 reverses the preference order. We label this
pattern the reflection effect.
Let us turn now to the implications of these data. First, note that the reflection
effect implies that risk aversion in the positive domain is accompanied by risk
seeking in the negative domain. In Problem 3′, for example, the majority of
subjects were willing to accept a risk of .80 to lose 4,000, in preference to a sure
loss of 3,000, although the gamble has a lower expected value. The occurrence of
risk seeking in choices between negative prospects was noted early by Markowitz
[29]. Williams [48] reported data where a translation of outcomes produces a
dramatic shift from risk aversion to risk seeking. For example, his subjects were
indifferent between (100, .65; – 100, .35) and (O), indicating risk aversion. They
were also indifferent between (-200, .80) and (-loo), indicating risk seeking. A
recent review by Fishburn and Kochenberger [14] documents the prevalence of
risk seeking in choices between negative prospects.
Second, recall that the preferences between the positive prospects in Table I are
inconsistent with expected utility theory. The preferences between the corresponding negative prospects also violate the expectation principle in the same
manner. For example, Problems 3′ and 4′, like Problems 3 and 4, demonstrate that
outcomes which are obtained with certainty are overweighted relative to
uncertain outcomes. In the positive domain, the certainty effect contributes to a
risk averse preference for a sure gain over a larger gain that is merely probable. In
the negative domain, the same effect leads to a risk seeking preference for a loss
PROSPECT THEORY
269
that is merely probable over a smaller loss that is certain. The same psychological
principle-the overweighting of certainty-favors risk aversion in the domain of
gains and risk seeking in the domain of losses.
Third, the reflection effect eliminates aversion for uncertainty or variability as
an explanation of the certainty effect. Consider, for example, the prevalent
preferences for (3,000) over (4,000, .80) and for (4,000, .20) over (3,000, .25). To
resolve this apparent inconsistency one could invoke the assumption that people
prefer prospects that have high expected value and small variance (see, e.g., Allais
[2]; Markowitz [30]; Tobin [41]). Since (3,000) has no variance while (4,000, .80)
has large variance, the former prospect could be chosen despite its lower expected
value. When the prospects are reduced, however, the difference in variance
between (3,000, .25) and (4,000, .20) may be insufficient to overcome the
difference in expected value. Because (-3,000) has both higher expected value
and lower variance than (-4,000, .80), this account entails that the sure loss
should be preferred, contrary to the data. Thus, our data are incompatible with the
notion that certainty is generally desirable. Rather, it appears that certainty
increases the aversiveness of losses as well as the desirability of gains.
Probabilistic Insurance
The prevalence of the purchase of insurance against both large and small losses
has been regarded by many as strong evidence for the concavity of the utility
function for money. Why otherwise would people spend so much money to
purchase insurance policies at a price that exceeds the expected actuarial cost?
However, an examination of the relative attractiveness of various forms of
insurance does not support the notion that the utility function for money is
concave everywhere. For example, people often prefer insurance programs that
offer limited coverage with low or zero deductible over comparable policies that
offer higher maximal coverage with higher deductibles-contrary to risk aversion
(see, e.g., Fuchs [16]). Another type of insurance problem in which people’s
responses are inconsistent with the concavity hypothesis may be called probabilistic insurance. To illustrate this concept, consider the following problem,
which was presented to 95 Stanford University students.
PROBLEM9: Suppose YOU consider the possibility of insuring some property
against damage, e.g., fire or theft. After examining the risks and the premium you
find that you have no clear preference between the options of purchasing
insurance or leaving the property uninsured.
It is then called to your attention that the insurance company offers a new
program called probabilistic insurance. In this program you pay half of the regular
premium. In case of damage, there is a 50 per cent chance that you pay the other
half of the premium and the insurance company covers all the losses; and there is a
50 per cent chance that you get back your insurance payment and suffer all the
losses. For example, if an accident occurs on an odd day of the month, you pay the
other half of the regular premium and your losses are covered; but if the accident
270
D. KAHNEMAN AND A. TVERSKY
occurs on an even day of the month, your insurance payment is refunded and your
losses are not covered.
Recall that the premium for full coverage is such that you find this insurance
barely worth its cost.
Under these circumstances, would you purchase probabilistic insurance:
N=95
Yes, No.
[20] [80]*
Although Problem 9 may appear contrived, it is worth noting that probabilistic
insurance represents many forms of protective action where one pays a certain
cost to reduce the probability of an undesirable event-without eliminating it
altogether. The installation of a burglar alarm, the replacement of old tires, and
the decision to stop smoking can all be viewed as probabilistic insurance.
The responses to Problem 9 and to several other variants of the same question
indicate that probabilistic insurance is generally unattractive. Apparently, reducing the probability of a loss from p to p/2 is less valuable than reducing the
probability of that loss from p/2 to 0.
In contrast to these data, expected utility theory (with a concave u) implies that
probabilistic insurance is superior to regular insurance. That is, if at asset position
w one is just willing to pay a premium y to insure against a probability p of losing
x, then one should definitely be willing to pay a smaller premium ry to reduce the
probability of losing x from p to (1- r)p, 0 < r < 1. Formally, if one is indifferent between (w -x, p; w, 1 -p) and (w - y), then one should prefer probabilistic insurance (w - x, (1 - r)p; w - y, rp; w - ry, 1 -p) over regular insurance (w - y). To prove this proposition, we show that implies Without loss of generality, we can set u(w - x ) = 0 and u(w) = 1. Hence, u(w y) = 1 -p, and we wish to show that which holds if and only if u is concave. This is a rather puzzling consequence of the risk aversion hypothesis of utility theory, because probabilistic insurance appears intuitively riskier than regular insurance, which entirely eliminates the element of risk. Evidently, the intuitive notion of risk is not adequately captured by the assumed concavity of the utility function for wealth. The aversion for probabilistic insurance is particularly intriguing because all insurance is, in a sense, probabilistic. The most avid buyer of insurance remains vulnerable to many financial and other risks which his policies do not cover. There appears to be a significant difference between probabilistic insurance and what may be called contingent insurance, which provides the certainty of coverage for a PROSPECT THEORY 27 1 specified type of risk. Compare, for example, probabilistic insurance against all forms of loss or damage to the contents of your home and contingent insurance that eliminates all risk of loss from theft, say, but does not cover other risks, e.g., fire. We conjecture that contingent insurance will be generally more attractive than probabilistic insurance when the probabilities of unprotected loss are equated. Thus, two prospects that are equivalent in probabilities and outcomes could have different values depending on their formulation. Several demonstrations of this general phenomenon are described in the next section. The Isolation Effect In order to simplify the choice between alternatives, people often disregard components that the alternatives share, and focus on the components that distinguish them (Tversky [44]). This approach to choice problems may produce inconsistent preferences, because a pair of prospects can be decomposed into common and distinctive components in more than one way, and different decompositions sometimes lead to different preferences. We refer to this phenomenon as the isolation effect. PROBLEM10: Consider the following two-stage game. In the first stage, there is a probability of .75 to end the game without winning anything, and a probability of .25 to move into the second stage. If you reach the second stage you have a choice between (4,000, .80) and (3,000). Your choice must be made before the game starts, i.e., before the outcome of the first stage is known. Note that in this game, one has a choice between .25 x .80 = .20 chance to win 4,000, and a .25 x 1.0 = .25 chance to win 3,000. Thus, in terms of final outcomes and probabilities one faces a choice between (4,000, .20) and (3,000, .25), as in Problem 4 above. However, the dominant preferences are different in the two problems. Of 141 subjects who answered Problem 10,78 per cent chose the latter prospect, contrary to the modal preference in Problem 4. Evidently, people ignored the first stage of the game, whose outcomes are shared by both prospects, and considered Problem 10 as a choice between (3,000) and (4,000, .80), as in Problem 3 above. The standard and the sequential formulations of Problem 4 are represented as decision trees in Figures 1 and 2, respectively. Following the usual convention, squares denote decision nodes and circles denote chance nodes. The essential difference between the two representations is in the location of the decision node. In the standard form (Figure I), the decision maker faces a choice between two risky prospects, whereas in the sequential form (Figure 2) he faces a choice between a risky and a riskless prospect. This is accomplished by introducing a dependency between the prospects without changing either probabilities or D. KAHNEMAN A N D A. TVERSKY FIGURE 1.-The FIGURE2.-The representation of Problem 4 as a decision tree (standard formulation). representation of Problem 10 as a decision tree (sequential formulation). outcomes. Specifically, the event 'not winning 3,000' is included in the event 'not winning 4,000' in the sequential formulation, while the two events are independent in the standard formulation. Thus, the outcome of winning 3,000 has a certainty advantage in the sequential formulation, which it does not have in the standard formulation. The reversal of preferences due to the dependency among events is particularly significant because it violates the basic supposition of a decision-theoretical analysis, that choices between prospects are determined solely by the probabilities of final states. It is easy to think of decision problems that are most naturally represented in one of the forms above rather than in the other. For example, the choice between two different risky ventures is likely to be viewed in the standard form. On the other hand, the following problem is most likely to be represented in the sequential form. One may invest money in a venture with some probability of losing one's capital if the venture fails, and with a choice between a fixed agreed return and a percentage of earnings if it succeeds. The isolation effect implies that the contingent certainty of the fixed return enhances the attractiveness of this option, relative to a risky venture with the same probabilities and outcomes. 273 PROSPECT THEORY The preceding problem illustrated how preferences may be altered ky different representations of probabilities. We now show how choices may be altered by varying the representation of outcomes. Consider the following problems, which were presented to two different groups of subjects. PROBLEM11: In addition to whatever you own, you have been given 1,000. You are now asked to choose between A: (1,000, .50), N=70 and B : (500). [16] [841* PROBLEM12: In addition to whatever you own, you have been given 2,000. You are now asked to choose between C: (-1,000, .50), N=68 and D : (-500). [69*] [311 The majority of subjects chose B in the first problem and C in the second. These preferences conform to the reflection effect observed in Table I, which exhibits risk aversion for positive prospects and risk seeking for negative ones. Note, however, that when viewed in terms of final states, the two choice problems are identical. Specifically, A = (2,000, .50; 1,000, .50) = C, and B = (1,500) = D. In fact, Problem 12 is obtained from Problem 11 by adding 1,000 to the initial bonus, and subtracting 1,000 from all outcomes. Evidently, the subjects did not integrate the bonus with the prospects. The bonus did not enter into the comparison of prospects because it was common to both options in each problem. The pattern of results observed in Problems 11 and 12 is clearly inconsistent with utility theory. In that theory, for example, the same utility is assigned to a wealth of $100,000, regardless of whether it was reached from a prior wealth of $95,000 or $105,000. Consequently, the choice between a total wealth of $100,000 and even chances to own $95,000 or $105,000 should be independent of whether one currently owns the smaller or the larger of these two amounts. With the added assumption of risk aversion, the theory entails that the certainty of owning $100,000 should always be preferred to the gamble. However, the responses to Problem 12 and to several of the previous questions suggest that this pattern will be obtained if the individual owns the smaller amount, but not if he owns the larger amount. The apparent neglect of a bonus that was common to both options in Problems 11 and 12 implies that the carriers of value or utility are changes of wealth, rather than final asset positions that include current wealth. This conclusion is the cornerstone of an alternative theory of risky cho'ice, whkh is described in the following sections. 274 D. KAHNEMAN AND A. TVERSKY 3. THEORY The preceding discussion reviewed several empirical effects which appear to invalidate expected utility theory as a descriptive model. The remainder of the paper presents an alternative account of individual decision making under risk, called prospect theory. The theory is developed for simple prospects with monetary outcomes and stated probabilities, but it can be extended to more involved choices. Prospect theory distinguishes two phases in the choice process: an early phase of editing and a subsequent phase of evaluation. The editing phase consists of a preliminary analysis of the offered prospects, which often yields a simpler representation of these prospects. In the second phase, the edited prospects are evaluated and the prospect of highest value is chosen. We next outline the editing phase, and develop a formal model of the evaluation phase. The function of the editing phase is to organize and reformulate the options so as to simplify subsequent evaluation and choice. Editing consists of the application of several operations that transform the outcomes and probabilities associated with the offered prospects. The major operations of the editing phase are described below. Coding. The evidence discussed in the previous section shows that people normally perceive outcomes as gains and losses, rather than as final states of wealth or welfare. Gains and losses, of course, are defined relative to some neutral reference point. The reference point usually corresponds to the current asset position, in which case gains and losses coincide with the actual amounts that are received or paid. However, the location of the reference point, and the consequent coding of outcomes as gains or losses, can be affected by the formulation of the offered prospects, and by the expectations of the decision maker. Combination. Prospects can sometimes be simplified by combining the probabilities associated with identical outcomes. For example, the prospect (200, .25; 200, .25) will be reduced to (200, .SO). and evaluated in this form. Segregation. Some prospects contain a riskless component that is segregated from the risky component in the editing phase. For example, the prospect (300, .80; 200, .20) is naturally decomposed into a sure gain of 200 and the risky prospect (100, .80). Similarly, the prospect (-400, .40; -100, .60) is readily seen to consist of a sure loss of 100 and of the prospect (-300, .40). The preceding operations are applied to each prospect separately. The following operation is applied to a set of two or more prospects. Cancellation. The essence of the isolation effects described earlier is the discarding of components that are shared by the offered prospects. Thus, our respondents apparently ignored the first stage of the sequential game presented in Problem 10, because this stage was common to both options, and they evaluated the prospects with respect to the results of the second stage (see Figure 2). Similarly, they neglected the common bonus that was added to the prospects in Problems 11 and 12. Another type of cancellation involves the discarding of common constituents, i.e., outcome-probability pairs. For example, the choice PROSPECT THEOKY 275 between (200, .20; 100, .50; -50, .30) and (200, .20; 150, .50; -100, .30) can be reduced by cancellation to a choice between (100, .50;-50, .30) and (150, S O ; -100, .30). Two additional operations that should be mentioned are simplification and the detection of dominance. The first refers to the simplification of prospects by rounding probabilities or outcomes. For example, the prospect (101, .49) is likely to be recoded as an even chance to win 100. A particularly important form of simplification involves the discarding of extremely unlikely outcomes. The second operation involves the scanning of offered prospects to detect dominated alternatives, which are rejected without further evaluation. Because the editing operations facilitate the task of decision, it is assumed that they are performed whenever possible. However, some editing operations either permit or prevent the application of others. For example, (500, .20; 101, .49) will appear to dominate (500, .15; 99, .5 1) if the second constituents of both prospects are simplified to (100, .50). The final edited prospects could, therefore, depend on the sequence of editing operations, which is likely to vary with the structure of the offered set and with the format of the display. A detailed study of this problem is beyond the scope of the present treatment. In this paper we discuss choice problems where it is reasonable to assume either that the original formulation of the prospects leaves no room for further editing, or that the edited prospects can be specified without ambiguity. Many anomalies of preference result from the editing of prospects. For example, the inconsistencies associated with the isolation effect result from the cancellation of common components. Some intransitivities of choice are explained by a simplification that eliminates small differences between prospects (see Tversky [43]). More generally, the preference order between prospects need not be invariant across contexts, because the same offered prospect could be edited in different ways depending on the context in which it appears. Following the editing phase, the decision maker is assumed to evaluate each of the edited prospects, and to choose the prospect of highest value. The overall value of an edited prospect, denoted V, is expressed in terms of two scales, .rr and v. The first scale, .rr, associates with each probability p a decision weight .rr(p), which reflects the impact of p on the over-all value of the prospect. However, .rr is not a probability measure, and it will be shown later that .rr(p)+.rr(l- p ) is typically less than unity. The second scale, v, assigns to each outcome x a number v ( x ) ,which reflects the subjective value of that outcome. Recall that outcomes are defined relative to a reference point, which serves as the zero point of the value scale. Hence, v measures the value of deviations from that reference point, i.e., gains and losses. The present formulation is concerned with simple prospects of the form (x,p; y, q ) , which have at most two non-zero outcomes. In such a prospect, one receives x with probability p, y with probability q, and nothing with probability 1 - p - q, where p + q s 1. An offered prospect is strictly positive if its outcomes are all positive, i.e., if x, y > 0 and p + q = 1; it is strictly negative if its outcomes
276
D. KAHNEMAN AND A. TVERSKY
are all negative. A prospect is regular if it is neither strictly positive nor strictly
negative.
The basic equation of the theory describes the manner in which .rr and v are
combined to determine the over-all value of regular prospects.
If (x,p; y , q ) i s a r e g u l a r p r o s p e c t ( i . e . , e i t h e r p + q < 1 , o r x > O > y , o r x ~ O ~
Y 1, then
(1)
V(x, p ; Y, q) = .rr(p)u(x)+ .rr(q)u(y)
where v(O)= O; ~ ( 0=)0, and ~ ( 1=) 1. As in utility theory, V is defined on
prospects, while v is defined on outcomes. The two scales coincide for sure
prospects, where V(x, 1.O) = V(x) = v (x).
Equation (1) generalizes expected utility theory by relaxing the expectation
principle. An axiomatic analysis of this representation is sketched in the Appendix, which describes conditions that ensure the existence of a unique .rr and a
ratio-scale v satisfying equation (1).
The evaluation of strictly positive and strictly negative prospects follows a
different rule. In the editing phase su…