using psychology theories in archival financial accounting...
TRANSCRIPT
Using Psychology Theories in Archival Financial Accounting Research
Lisa Koonce Department of Accounting
McCombs School of Business CBA 4M.202
The University of Texas Austin, TX 78712
Molly Mercer Goizueta Business School
Emory University 1300 Clifton Road Atlanta, GA 30322
January 5, 2005
We appreciate the helpful comments and suggestions of Jan Barton, Sarah Bonner, Shane Dikolli, Jeffrey Hales, Eric Hirst, Steve Kaplan, Jay Koehler, Grace Pownall, Brad Schafer, Greg Waymire, and workshop participants at the 2002 Accounting, Behavior, & Organizations conference.
Using Psychology Theories in Archival
Financial Accounting Research
Abstract Psychologists have studied human behavior for over a century and, as a result, have developed a robust set of theories regarding how people behave. Most financial accounting issues deal with matters of human behavior, such as the judgments and decisions of managers, investors, analysts, and auditors. Consequently, psychology offers a rich pool of theories from which financial accounting researchers can draw to motivate hypotheses and interpret results. Despite this, archival accounting researchers traditionally have relied almost solely on theories based in financial economics. We argue that two major obstacles to the use of psychology theories by archival researchers has been a lack of awareness about the theories that are available and when their use would be most productive. Our paper attempts to bridge this gap. Specifically, we describe a number of psychology theories that are applicable to financial accounting issues, lay out the circumstances where they may be especially useful to archival researchers, and provide a number of specific examples of how psychology theories provide new insights about financial accounting issues.
1
1. Introduction
Virtually all financial accounting issues involve decision-making. For example, managers
decide how to craft their voluntary disclosures, auditors decide whether to issue unqualified audit
opinions, investors decide whether to buy or sell stocks, and sell-side analysts decide how to
interpret financial reports. Psychologists have studied human decision making for over 100
years and have identified a number of robust theories about how people behave. Thus,
psychology research can provide a wealth of insights about the behaviors of managers, auditors,
investors, and analysts. Experimental accounting researchers have long recognized the relevance
of psychology theories to financial accounting issues (Libby et al. 2001); however, archival
accounting researchers have rarely tapped into these theories to make predictions or interpret
results.1 The purpose of our paper is to argue that psychology theories can enhance the insights
offered by archival researchers studying financial accounting issues. Testing psychology-based
predictions with archival data is important because these tests provide evidence on whether/when
psychology-based predictions hold under real-world market conditions.
The dearth of psychology-based archival research is readily apparent when looking at recent
publications in top-tier accounting journals. We surveyed all empirical financial accounting
papers published in The Accounting Review, Contemporary Accounting Research, Journal of
Accounting and Economics, Journal of Accounting Research, and Review of Accounting Studies
over the last ten years. Each paper was classified according to its methodology (archival or
experimental) and the primary type of theory used to make predictions and interpret results
(economic or psychology). The results of this analysis are reported in Figure 1. We find that
approximately 71 percent of experimental financial accounting papers rely on psychology
theories, compared with only 2 percent of archival financial accounting papers.
2
[Insert Figure 1]
There are several reasons for the dearth of psychology-based archival studies in accounting.
First, many archival researchers do not have training in psychology.2 This lack of training is
due, in part, to the structure of accounting doctoral programs. Accounting doctoral students
typically choose one of two, largely non-overlapping, fields of specialization—economics or
psychology. This choice often is determined by the type of methodology the student expects to
use in his or her research; students who wish to work with archival data take coursework in
finance and economics, while those desiring to be experimentalists take psychology coursework.
Thus, the typical financial accounting researcher tends to be either knowledgeable about archival
methods but not psychology theories or knowledgeable about psychology theories but not
archival methods.3
The second obstacle, though related to the first, is more philosophical in nature. Traditional
economic theories, which are the basis of most archival research, often rely on assumptions that
deem psychology theories irrelevant. For example, economics-based research traditionally has
been based on the assumption that people behave in a manner consistent with normative models
and that even if they do not behave rationally, markets forces will eliminate individual irra-
tionality (Camerer 1987; Kothari 2001). Thus, archival accounting researchers, who are trained
in economics, often judge as irrelevant the individual-level behaviors observed by psychologists.
Impediments caused by this second obstacle have been fading to some extent as researchers
in economics and finance have begun to question whether markets eliminate all individual-level
judgment biases (Schleifer 2000; Daniel et al. 2002). For example, Camerer and Hogarth (1999)
recently performed a meta-analysis on 74 incentive-related studies. Based on these studies, they
concluded that irrational behaviors often persist even in the presence of high incentives such as
3
those found in the financial markets (see also Kachelmeier and Shehata 1992). In fact, a greater
acceptance of the notion that irrational behaviors can persist in market settings has led to the
emergence of new psychology-based subfields within economics and finance, aptly named
behavioral economics and behavioral finance (Lee 2001; Thaler 1999). Thus, philosophical
obstacles to the use of psychology theories in archival accounting research have lessened.
Indeed, identifying the conditions where psychology theories add to our understanding of market
behaviors is a fruitful area for future research (Lee 2001).
Our paper provides a number of examples of how psychology theories can help archival
researchers by allowing for different and/or more specific predictions in many areas of financial
accounting research. Our goal is to demonstrate how psychology theories can be useful to
archival researchers and, thus, prompt archival research on both these and other financial
accounting issues. To accomplish this, we first identify and briefly review theories from the two
areas of psychology that we judge as particularly relevant to financial accounting researchers –
cognitive psychology and social psychology. Pinpointing the relevant areas is necessary, as
psychology is a very large and diverse field with many areas not directly applicable to
accounting. Theories from cognitive and social psychology address a number of issues relevant
to accounting, such as how people search for new information (e.g., Koehler 1996), make
inferences (e.g., Kelley 1972), evaluate risky alternatives (e.g., Kahneman and Tversky 1979),
and develop expertise (e.g., Feltovich et al. 1997). This review provides an overview of the
types of insights that can be drawn from psychology research.
We then draw on this discussion and outline circumstances where psychology-based theories
can add to our understanding of financial accounting issues above and beyond that provided by
economic theories. Identifying these situations is important as we are not advocating that
4
psychology replace economics. Rather, we believe that economic and psychology theories
should be treated as complementary, and the most suitable theory or theories be used. One
circumstance where psychology can add to our understanding is when it makes different
predictions than economics. As noted earlier, economic theories often predict that individuals
will act rationally in accordance with normative models such as Bayes’ theorem or expected
utility theory. Psychology research, in contrast, tends to be descriptive and does not assume that
people behave in accordance with normative models. Thus, psychology theories sometimes lead
to different predictions than those from economics and may help explain seemingly anomalous
(from an economics perspective) market responses to financial accounting information. Further,
because the manner in which individual behavior translates into collective market behavior is not
well understood (Lee 2001), archival researchers are in a unique position to contribute to the our
understanding of when the systematic individual behaviors documented by psychologists are
reflected in market settings (see, for example, Burgstahler and Dichev 1997).
We also argue in this paper that psychology theories may be useful to archival researchers
even when they do not contradict theories from economics. That is, psychology theories also
allow for more-specific predictions than economics in some areas. One of the hallmarks of
psychology research is that it aims to describe the process by which judgments and decisions are
formulated. Economic theory, in contrast, is frequently silent regarding the process underlying
judgments and decisions. Put another way, whereas economists tend to focus on predicting the
equilibrium behavior that results under specified conditions, psychologists generally emphasize
understanding the causal mechanisms or processes underlying behavior that could lead to that
equilibrium. For example, while economic theories can describe the equilibrium prices that
result if managers have different levels of reporting reputation, they provide less insight into how
5
managers develop reputation over time. Psychology theories, in contrast, suggest a number of
key variables that influence how managers develop their reputation over time. Given the quest to
understand “how, when, and why prices adjust (or fail to adjust) to information” (Lee 2001, 237,
emphasis added), theories from psychology are likely to be highly relevant as they may provide
possible explanations for observed market responses.
This paper is organized as follows. Section 2 describes some key findings and theories from
two areas of psychology that are particularly relevant to financial accounting issues. Our
discussion in this section relies largely on citations to the original psychology research where
these theories were developed in order to provide the best description of the basic psychological
findings without introducing complexities from a particular applied setting.4 Section 3
articulates circumstances where archival researchers can productively use these theories in their
research, providing a number of specific examples across a variety of financial accounting
contexts. Section 4 summarizes and discusses limitations to the use of psychology theories in
archival financial accounting research.
2. Pertinent psychology theories
Because of the large size of the psychology field, various subfields have developed.
Although each subfield shares psychology’s common goal of understanding mind and behavior,
these subfields vary widely in terms of the specific topics they address. Two subfields whose
topical coverage and level of analysis make them particularly relevant for financial accounting
researchers are cognitive psychology, which investigates how people think, and social
psychology, which focuses on how features of the social environment influence how people think
and behave.
6
Cognitive psychology Judgment and decision making
Judgment and decision making (JDM) is a branch of cognitive psychology that should be
especially interesting to archival researchers because it lies at the crossroads of cognitive
psychology and economics.5 Much of cognitive psychology focuses on providing descriptive
evidence of how people think but does not address whether these behaviors are consistent with
normative models of how they should think. In contrast, JDM research often explicitly compares
people’s actual decisions to economics-based models of normative decisions. JDM research also
differentiates itself from other research in cognitive psychology because of its focus on readily
observable outputs. This focus makes JDM research highly relevant to applied disciplines such
as accounting. We describe JDM’s insights on two key issues: 1) how people make decisions
under conditions of uncertainty, and 2) how problem framing affects decision-making.
Decision-making under uncertainty. Until the mid-1900’s, expected utility (EU) theory
from economics was considered to be both a normative and descriptive theory of decision
making under uncertainty (von Neumann and Morgenstern 1947; Savage 1954). According to
EU, people choose between risky alternatives based on the utility derived from potential
outcomes, where each possible outcome is weighted by its probability of occurrence. For
example, assume that x and y represent two potential outcomes under Alternative A, where u is
the utility derived from the outcome, and p is the probability that the outcome will occur. EU
states that the overall utility derived from Alternative A is u(x) * p(x) + u(y) * p(y). If the utility
of A, computed in this manner, is higher than the utility of other alternatives, then the decision
maker will prefer and choose A.6
7
In the latter half of the 1900’s, evidence began to mount that decision makers sometimes
violate the predictions of EU in predictable ways (e.g., Allais 1953; Ellsberg 1961; Tversky
1969; Lichtenstein and Slovic 1971; Rabin 1998), leading researchers to question the descriptive
adequacy of EU. For example, Markowitz (1952) challenged EU’s assumption that people
evaluate the outcomes of risky decisions in terms of final wealth states. He proposed that people
encode financial outcomes as gains and losses rather than as final wealth states, and that this
process must be a focal point in any descriptive model of decision-making. In 1979, Kahneman
and Tversky provided such a model, which they called Prospect Theory. This theory accounted
for many previously detected EU violations. According to Prospect Theory, risk attitudes are
determined jointly by a context-dependent value function (v) and a probability weighting
function (π). In prospect theory, the value derived from Alternative A described above would be
represented as v(x) * π(p(x)) + v(y) * π(p(y)).
Prospect theory’s v, shown in Panel A of Figure 2, differs from the utility function in several
important ways. First, v has a different shape for gains and losses. For gains, v is concave and
has an appearance similar to that provided by the expected utility models. However, v is convex
for losses, suggesting that the marginal value of both gains and losses decreases with their size.
Second, v is steeper for losses than for gains. This slope differential captures the idea that the
amount of displeasure caused by a loss is roughly twice the amount of pleasure caused by an
equal-sized gain (Tversky and Kahneman 1991). The idea that losses are more displeasurable
than the equivalent gain is pleasurable is commonly referred to as “loss aversion.” Third,
whereas utility is usually defined in terms of net wealth, prospect theory defines v in terms of
gains and losses from reference points. This definition suggests that one cannot determine
8
whether an outcome will be perceived as a gain or loss unless until one has some understanding
of the points of reference used by decision makers.
[Insert Figure 2] The probability weighting function (π) in prospect theory also differs from the probabilities
used in EU. Whereas expected utility theory assumes that people properly assess the
probabilities of various outcomes, the decision weighting function assumes that people
overweight small probabilities and underweight moderate and large probabilities (see Panel B of
Figure 2). The decision weighting function helps explain a number of previously anomalous
findings. For example, the idea that people overweight small probabilities explains people’s
willingness to buy lottery tickets and their preferences for regular, rather than probabilistic,
insurance (Kahneman and Tversky 1979).
Framing effects. Standard economic theory presumes that the way in which a situation or set
of options is described does not and should not affect the preferences that people have. This
“invariance” principle (Tversky and Kahneman 1986, 253) is a “fundamental element of
rationality” (Arrow 1982, 6). Invariance is an essential component of the normative economic
theory because violations of this principle imply that people do not have stable preferences. In
other words, if variations of form that do not affect actual outcomes impact the choices that
people make, then one cannot maintain that people have a stable preference ordering for the
available outcomes.
An overwhelming amount of evidence in psychology suggests that the invariance principle is
as descriptively invalid as it is normatively compelling. That is, psychology studies show that
the ways in which economically equivalent options are presented have a large impact on the
9
choices made by decision makers. Consider the now classic Asian Disease problem that first
appeared in Tversky and Kahneman (1981):
Imagine that the U.S. is preparing for the outbreak of an unusual Asian disease, which is expected to kill 600 people. Two alternative programs to combat the disease have been proposed. Assume that the exact scientific estimate of the consequences of the programs are as follows: If Program A is adopted, 200 people will be saved. If Program B is adopted, there is 1/3 probability that 600 people will be saved, and 2/3 probability that no people will be saved.
When people are asked to choose between Programs A and B, most select Program A. However,
when Tversky and Kahneman re-frame the program consequences in terms of the number of
people who will die (rather than the number who will live), they find dramatically different
results. The “die” version of the problem replaces the last two sentences of the version above as
follows:
If Program C is adopted, 400 people will die. If Program D is adopted, there is 1/3 probability that nobody will die, and 2/3 probability that 600 people will die. Looking across the two versions of the Asian Disease problem, Programs A and C have
equivalent consequences, as do Programs B and D. Consequently, standard economic theory
predicts that respondents will choose options A and C (and options B and D) in identical
proportions. Such a prediction is required by the principle of invariance. Furthermore, by the
principle of decreasing marginal utility, economic theory predicts that respondents will tend to
select the risk-averse option A over B and, correspondingly, option C over D.
In contrast, prospect theory (Tversky and Kahneman, 1981) draws on the various
psychological principles described earlier (e.g., gains and losses relative to flexible reference
points and different risk attitudes in the gain and loss domains) to offer different predictions.
Prospect theory predicts that framing the problem in terms of lives saved rather than lives lost
will cause people to think about the problem in terms of gains (i.e., number who will live) rather
10
than losses (i.e., number who will die). Because the natural reference point in the “live” frame is
zero lives saved, options A and B both reside in the domain of gains and prospect theory, like
expected utility theory, predicts that people will tend to favor the risk averse option (A).
However, because the natural reference point in the “die” frame is zero lives lost, options C and
D in the second version of the problem both reside in the domain of losses. If people tend to be
risk-seeking in the domain of losses as prospect theory predicts, then people will be more likely
to favor the riskier option (D). This is precisely the pattern of results that Tversky and
Kahneman observe. They find that whereas only 28% of people prefer the risky option (A) in
the “live” frame, 78% prefer the reworded – but otherwise identical – risky option (D) in the
“die” frame. This result violates the principle of invariance and provides evidence that the way
in which identical options are presented can exert a significant influence on people’s choices.
Determinants and consequences of expertise
Another area of cognitive psychology that may be particularly relevant to accounting
researchers is the literature on the determinants and consequences of expertise. Most economic
theories are silent as to how people acquire expertise. In contrast, theories from cognitive
psychology not only outline how people become experts but also identify the tasks in which
expertise is most likely to affect performance. Below we separately consider the psychology
literature on the determinants and consequences of expertise.
Determinants of expertise. How does one become an expert? The evidence points to
several inter-related factors. Aptitude appears to play an important role (Sternberg and
Grigorenka 2003). Not surprisingly, experience plays an important role as well (Seifert et al.
1997). Importantly, the effects of aptitude and experience on expertise appear to be mediated by
knowledge (Hunter 1983; Schmidt et al. 1986; Libby and Luft 1993). In other words, those who
11
practice a skill or who have natural aptitude in an area are more likely to increase their relevant
knowledge base; this knowledge, in turn, promotes the development of expertise. One
implication of this research is that experience in and of itself will not promote expertise unless
the experience expands the individual’s knowledge base.
Whereas an individual’s aptitude, knowledge, and experience provide cues to his or her level
of expertise, these person-specific features alone may not provide a secure basis for predicting
performance. It is well-known in psychology that human behavior is best understood when the
features of the person are examined in light of the features of the situation that the person
confronts. We consider two task features – task complexity and task feedback – that have been
shown to play important roles in whether experts will be able to translate their expertise into
superior performance.
The existing research suggests that if tasks are extremely easy or extremely difficult (in terms
of the base rate probability of finding a successful solution), there may be no obvious benefit to
expertise (Bonner 1994). But for non-extreme degrees of complexity, the benefits of expertise
generally increase as the difficulty of the task increases. The type of feedback that people
receive after making predictions, judgments, and choices also provides an indication of the likely
value of expertise. In environments where feedback is timely, accurate and complete, people can
learn from their mistakes and make better judgments in the future (Einhorn 1980). Consistent
with this principle, experts in domains with complete, timely feedback, such as sports betting or
weather forecasting, tend show remarkable accuracy in their judgments (Dowie 1976; Murphy
and Winkler 1984; Stewart et al. 1997).7 However, experts in domains with lower quality
feedback, such as physicians, attorneys, and financial analysts, are significantly less accurate in
their judgments and decisions (Koehler et al. 2002).
12
Consequences of expertise. According to classic studies in cognitive psychology, experts
are better than novices at organizing information in memory, recognizing underlying patterns in
data, and matching stimuli to their relevant categories (Chase and Simon 1973; Chi et al. 1982;
Glaser 1990; Simon 1979). For example, Chi et al. (1981) report that novices classify physics
problems according to superficial similarities whereas physics experts classify problems
according to the sophisticated principles that are required to solve the problems. Not
surprisingly, these processing differences are generally reflected in task performance; experts
have demonstrated greater proficiency than novices across a range of tasks (Feltovich et al
1997).8 Experts often demonstrate not just better performance, but also more efficient
performance. That is, studies also have pointed to the importance of “automaticity” in expertise,
wherein experts are able to perform tasks automatically, with significantly lower cognitive effort
(Shiffrin and Schneider 1977).
Despite experts’ performance advantages, a number of studies have shown that experts fall
prey to some of the same judgment biases as non-experts (e.g., Neale and Northcraft 1986;
Redelmeier and Shafir 1995; Camerer et al. 1997). The fallibility of experts is surprising to
some economists, who presume that if an expert commits the same judgment errors as a novice,
the resulting poor performance will drive the expert from the marketplace. But there are reasons
to believe that the cognitive workings of experts are sometimes quite similar to those of non-
experts. For example, experts are just as likely to be overconfident (Einhorn and Hogarth 1978),
make overly optimistic estimates (Hogarth and Makridakis 1981; Cadsby 2000; Glare et al.
2003), and underweight base rates (Joyce and Biddle 1981; Eddy 1982; Kida 1984; Balla et al.
1985). In fact, studies show that under certain circumstances, experts’ experience and
knowledge causes them to perform worse than novices. Specifically, experts have been shown to
13
underperform novice decision makers in unusual task conditions because novices are less likely
to be locked into specific problem-solving techniques (Frensch and Sternberg 1989).9
Summary
Cognitive psychology is the study of thought. Researchers in this subfield study a broad
range of relevant issues, including how people attend to and perceive information, learn, develop
expertise, reason, solve problems, and make choices under uncertainty. Because these issues are
fundamental to the financial reporting process, cognitive psychology contains a wealth of
insights for archival researchers attempting to unravel financial reporting issues. We now turn to
social psychology, another subfield that is especially relevant to financial accounting researchers.
Social psychology
Social psychology is a branch of psychology that attempts to “understand and explain how
thoughts, feelings and behaviours of individuals are influenced by the actual, imagined or
implied presence of others” (Allport 1954). Thus, social psychologists study many of the same
general issues as cognitive psychologists, but focus on how features of the social environment
influence people’s judgments and decisions. For example, whereas cognitive psychologists
might examine how individuals combine quantitative inputs to arrive at an overall forecast,
social psychologists would be more interested in the role that multiple forecasters and/or
different incentives schemes play in the forecasting process. In the sections below, we describe
several classic theories from social psychology that describe, (1) how people draw inferences
from others’ behaviors, and (2) how people search for and process new information when
updating their beliefs.
14
Inferences about others’ behaviors
A fundamental issue in social psychology concerns how people explain the causes of other
people’s behaviors. These explanations or “attributions” are important because people use them
to predict future behaviors. Heider (1958) proposed a simple and useful dichotomy for these
attributions: internal and external. Internal attributions focus on a personality trait or underlying
dispositional feature of an individual. External attributions focus on situational causes for
behavior. As research in this area developed, it became clear that people tend to underrate the
explanatory power of situational explanations relative to dispositional influences (Ross 1977).
For example, people are more likely to assume that a person makes a donation because he or she
is dispositionally kind (internal attribution) rather than because situational forces were exerted on
the donor (external attribution). Interestingly, people are much more sensitive to situational
forces where their own behavior is concerned (Jones and Nisbett 1972) or when their attention is
directed to the self-interest of another individual. The idea that inferences are influenced by
incentives is central to correspondent inference theory, one of the major attribution theories.
Correspondent inference theory. The theory of correspondent inference was developed by
social psychologists to help explain when people will attribute the behavior of another person to
an internal, dispositional trait rather than to external, situational forces (Jones and Davis 1965).
According to the theory, people are especially likely to attribute behavior to internal
characteristics when the behavior is voluntary, unexpected, and does not have obvious benefits to
the individual being observed. As we discuss in Section 3, one implication of this theory is that
people may be able to influence their reputations by creating a perception that their positive
actions are voluntary, surprising, or without personal benefit, whereas their negative actions are
coerced and/or unsurprising.
15
Covariation theory. Kelly (1967) developed an alternative attribution theory – covariation
theory – that focuses on how other information influences our inferences about an observed
behavior. According to covariation theory, people consider three potential covariates when
deciding whether to attribute observed behavior to internal or external forces: (a) consensus
information (i.e., how do others behave in this situation?), (b) consistency (i.e., does this person
always act this way in this situation?), and (c) distinctiveness (i.e., does this person act similarly
in other situations?). Covariation theory suggests that when consensus is low, consistency is
high, and distinctiveness is low, an internal attribution is most likely. Conversely, when
consensus is high, consistency is high, and distinctiveness is high, an external attribution is most
likely. Like correspondent inference theory, covariation theory implies that individuals may be
able to manipulate others’ attributions (internal or external) for their actions.
Persuasion theories. Finally, researchers in social psychology have used attribution theory
to make predictions about when messages will be persuasive. Attribution theories suggest that
people will find messages that are inconsistent with the source’s incentives to be more persuasive
than messages that are consistent with the source’s incentives. Specifically, when a message is
consistent with the incentives of the source, we are more likely to attribute the message to those
incentives and less likely to attribute the message to the source’s true beliefs. Conversely, when
a message is inconsistent with the source’s incentives, we are more likely to attribute the
message to the source’s true beliefs. Consequently, incentive-inconsistent messages tend to be
more persuasive than incentive-consistent messages (Eagly et al. 1978; Birnbaum and Stegner
1979).
16
Belief updating Social psychologists also study how and when people update their initial beliefs. Unlike
economists, psychologists do not necessarily assume that people update their beliefs in a rational
Bayesian manner. Indeed, research in social psychology suggests a number of systematic biases
in belief updating. For example, research in this area suggests that once formed, the beliefs that
people hold are often surprisingly resistant to change. Various factors appear to underlie this
phenomenon. First, people are motivated to have consistent attitudes. Consequently, when
presented with evidence contrary to their beliefs about a person or object, people experience
“cognitive dissonance” (Festinger 1957). Dissonance refers to an unpleasant state of tension that
arises when two or more thoughts, cognitions, or evidentiary items appear to be in conflict.
People generally try to relieve dissonance by rationalizing away the inconsistency or by altering
their attitudes to bring the cognitions into alignment. However, as discussed below, psychology
studies indicate that the former strategy is more likely because pre-existing beliefs are often quite
resistant to change.
Biased information search. Research within psychology suggests that beliefs are often
resistant to change as a result of biased information search (Klayman and Ha 1987).
“Confirmation bias” is one of the most frequently studied search biases. This bias refers to a
tendency for decision makers to seek information that is consistent with their pre-existing beliefs.
Importantly, confirmation bias appears even in situations where there is no obvious incentive or
motivation to believe the favored hypothesis. For example, Doherty et al. (1979) find that when
participants in a laboratory experiment are given the opportunity to search for information
pertinent to determining which of two islands produced a hypothetical artifact, the participants
tend to request information that confirms – rather than properly tests – the validity of their initial
17
beliefs. In other words, even though the participants had no external motivation to care one way
or the other about where the artifacts were produced, they consistently used confirmatory search
strategies to investigate the issue.
Biased processing of disconfirming information. When evidence that tends to disconfirm
a favored hypothesis does appear, people are more critical of the evidence than they are of
confirmatory evidence (Edwards and Smith 1996). This tendency appears even among people
who are trained in the scientific method. Koehler (1993) finds that scientists rate evidence that
supports favored hypotheses to be methodologically superior to identical evidence that
disconfirms those hypotheses. He also demonstrates that the tendency to rate belief-confirming
evidence as being of higher quality is greatest among scientists whose prior beliefs are strongest.
Evidence for the cognitive origins of this information processing bias appears in studies that
show that people have more trouble retrieving and accurately recalling disconfirming evidence
than confirming evidence (Koehler 1993, 29).
Belief perseverance and polarization. A byproduct of information search and processing
biases is that people’s initial beliefs tend to be sticky. If people tend not to search for and find
belief-disconfirming evidence and if they tend to discount such evidence when it arises, then
people are unlikely to be persuaded to discard their pre-existing beliefs. Lord et al. (1979) show
that one paradoxical implication of biased information processing is belief polarization. If
individuals on both sides of a dispute discount evidence that opposes their initial beliefs while
bolstering the quality of evidence that supports those beliefs, they are likely to become even
more convinced about the veracity of their initial beliefs following exposure to a mixed set of
evidence. In other words, people’s beliefs may diverge rather than converge following exposure
to an identical set of evidence.
18
Psychologists have identified a related class of belief perseverance phenomena in a series of
studies that show that people’s initial beliefs sometimes persist even after the entire evidentiary
basis for those beliefs has been discredited. For example, Ross et al. (1975) provided people
with a set of suicide notes and asked them to identify which were real and which were not.
Along the way, the participants were given false feedback indicating their performance in this
task. The participants were later informed about that the feedback they received was false.
Nevertheless, when asked to predict how well they actually performed on the task, participants
continued to rely on the information contained in the false feedback. That is, people who were
falsely told that they were very good at determining suicide note authenticity continued to
believe that they performed well at this task, whereas people who were falsely told that they
were very bad at the task believed that they performed poorly.
It is hard to imagine a rational justification for belief perseverance when the evidentiary basis
for the belief is completely and convincingly discredited. In some cases, a motivational
explanation has merit. That is, people sometimes benefit from maintaining a consistent belief
and are harmed by switching beliefs. For example, an academic researcher who stakes his
reputation on the veracity of a theory obviously has good reason to resist evidentiary challenges
to that theory. But why might a thoroughly discredited belief persevere among the participants
in Ross et al. (1975), who have no motivation to maintain a belief in their own competence or
incompetence in a suicide note authenticity task? Psychology theory offers an intriguing
possibility. When people receive information pertinent to a belief, they integrate this
information with other beliefs in a way that helps maintain and confirm those beliefs. Thus,
when someone is told that he performed poorly in the suicide note authenticity study, he might
view this as consistent with other things he knows about himself such as “I can’t tell when my
19
girlfriend is being truthful or not” or “I don’t know much about why people commit suicide.”
Such confirmatory thoughts linger even after the feedback that triggered those thoughts has been
discredited. The result is that the discredited false feedback continues to influence beliefs, albeit
indirectly, through the confirmatory thoughts that the feedback triggered.
Summary
In sum, social psychologists investigate how people’s thoughts and behaviors are influenced
by features of their social environment. Researchers in this area have developed and validated a
number of theories regarding how the presence of others influences people’s thoughts and
behaviors. Social psychology’s focus on multi-person environments makes it highly relevant to
many financial accounting issues.
3. Applying psychology theories to archival financial accounting research We argue that there are two primary reasons why psychology theories, such as those from
cognitive and social psychology, are relevant to archival researchers. First, cognitive and social
psychology theories often allow for different predictions than could be made using economic
theories alone. As we note in Section 2, psychology theories generally do not rely on the strict
rationality assumptions underlying most theories from economics.10 Consequently, these
theories often lead to different predictions about how people behave (cf. Kahneman and Tversky
1979; Rabin 1998). Drawing on this notion, we identify a number of new predictions that can be
made using these theories. Specifically, we provide concrete examples of how cognitive and
social psychology theories allow for different predictions than economics for three important
financial accounting issues—namely, why firms manage earnings, how disclosure format affects
investors’ use of financial information, and how financial analysts’ incentives influence their
research reports.
20
In this section of the paper, we also argue that psychology theories may be useful to archival
researchers even when they do not contradict theories from economics. Whereas economists
tend to focus on predicting the equilibrium behavior that result under specified conditions,
psychologists generally emphasize understanding the causal mechanisms or processes underlying
behavior. Because of this difference in approach, psychology theories also allow for more-
specific predictions about how people behave. For example, while economic theories can
describe the equilibrium prices that result if managers have different levels of reporting
reputation, they provide less insight into how managers develop reputation over time. Similarly,
economic theories predict the equilibrium forecasts that result if analysts differ in their levels of
expertise, but not how analysts develop expertise. Accordingly, in the second half of this
section, we provide three specific financial accounting examples of how the greater specificity of
process provided by psychology theories helps fill these gaps.
Financial accounting examples where psychology theories lead to different predictions
Understanding earnings management
A large body of archival research examines how and why firms manage their financial results
(Healy and Wahlen 1999; Fields et al. 2001). Predictions from these studies are usually based on
economics-based arguments where earnings management results from managers’ rational
choices for a given set of constraints. For example, some studies predict that managers
manipulate earnings to maximize their compensation under the terms of their bonus plan (Healy
1985). Other studies suggest that firms manage earnings to avoid violating the terms of their
debt agreements (Defond and Jiambalvo 1994) or to avoid political scrutiny (Key 1997).
Prospect theory, a well-developed JDM theory about decision making under uncertainty,
suggests some explanations for earnings management that would not ordinarily be considered by
21
archival researchers. As explained in Section 2, prospect theory indicates that people judge
outcomes, in part, by a myriad of considerations that are based on psychological, not economic,
principles. These considerations may translate into a preference for certain earnings patterns
which, in turn, may help explain a number of alleged earnings management behaviors: 1)
attempts to avoid reporting small losses, 2) a willingness to take “big baths,” 3) the creation of
“cookie jar reserves,” and 4) attempts to influence the points of comparison that investors use to
evaluate reported earnings.
Prospect theory provides a descriptive (rather than normative) theory of decision making
under uncertainty. To the extent that firms are sensitive to the idea that investors may have
preferences for particular earnings patterns, the shape of the prospect theory value function
implies that firms will manage earnings to avoid small earnings losses (Burgstahler and Dichev
1997). Recall that the value function is steeper in the loss domain than in the gain domain (see
Figure 2). This feature implies that the pain associated with a loss is greater than the pleasure
associated with an equal-sized gain. The steepness of the value function in the loss domain
implies that investors will experience relatively great disutility from small reported losses. In
addition, the concavity of the value function in the gain domain implies that investors will
experience diminishing marginal utility from additional reported gains. Taken together, these
features suggest that, ceteris paribus, investors will prefer to invest in companies that report a
series of small gains rather than companies that include some large gains and some small losses.
To illustrate, consider two firms that are identical in all respects other than the pattern of their
reported earnings. Both firms earn 11 cents per year across two years. However, Firm A earns 2
cents per share in Year 1 and 9 cents per share in Year 2, whereas Firm B experiences a 2-cent
loss in Year 1 and a 13-cent gain in Year 2. If Firm B does not manage earnings, prospect theory
22
predicts that, all else equal, investors will prefer to invest in Firm A. This preference arises
because the subjective value investors derive from the combination of +2 cents and +11 cents is
higher than the combination of -2 cents and +13 cents. Prospect theory also predicts that if Firm
B “borrows” 4 cents in earnings from Year 2 for placement in Year 1, the preference investors
would otherwise show for Firm A will disappear. In short, prospect theory provides an
explanation for why managers are likely to manage earnings to avoid reporting small losses.
Prospect theory also offers an explanation for why firms take “big baths” once the prospect
of reporting a gain is untenable. When investors view earnings within the domain of losses, their
value functions for further gains and losses are apt to be convex, rather than concave, in shape
(see Figure 2). This value function implies that the incremental disutility investors are likely to
experience after learning about additional losses is relatively small. Therefore, if a firm must
report a loss in a given period, prospect theory predicts that knowledgeable firms will manage
earnings to include expected future losses in the current period as well. Furthermore, prospect
theory predicts that firms with sizable losses in the current period will try to shift any gains from
the current period into future periods. According to the shape of the value function, the small
decrease in disutility associated with reducing a very large loss to a slightly small loss in the
current period is outweighed by the increase in positive utility associated with reporting a small
gain in a future period. Thus, prospect theory suggests that firms that are unable to report
positive earnings will ‘take a bath’ by (a) including all possible losses, and (b) excluding all
possible gains from current year earnings.
Prospect theory also makes predictions about how firms will manage earnings in especially
profitable times. The concavity of the value function for gains implies that investors will prefer
to see gains broken out and reported in separate periods rather than reported all at once in a
23
single period. Consequently, if a firm’s earnings significantly exceed expectations in a particular
year, prospect theory suggests that the knowledgeable firms will manage earnings (i.e., create
“cookie jar reserves”) to reserve part of the gain for a subsequent year.
Finally, prospect theory predicts that investors will evaluate a company’s reported results
relative to some reference point. Results that are greater than the reference point will be
experienced as ‘gains,’ and results that are less than the reference point will be experienced as
‘losses’ (Tversky and Kahneman 1981). The notion that people define gains and losses relative
to unstable reference points rather than in terms of absolute wealth states leads to a set of
predictions about investor behavior that might not be anticipated by conventional economic
theories.
Investors appear to use a variety of natural reference points, including earnings in a prior
year, the consensus analyst forecast number, and zero (Degeorge et al. 1999). Consequently,
prospect theory predicts that savvy managers will try to report results that exceed these reference
points. Some managers may try to reach this goal by managing the reported earnings number
(Burgstahler and Dichev 1997; Degeorge et al. 1999). Other managers may try to report results
that exceed investors’ reference points by manipulating the reference points themselves. For
example, they may guide analysts’ forecasts downward (Matsumoto 2002) or provide strategic
presentations of prior year results (Schrand and Walther 2000).
Although prospect theory from psychology implies implies a number of specific predictions
about when and why firms might manage earnings,11 economics-based contracting theories also
predict some of these earnings management behaviors, albeit by different mechanisms (Watts
and Zimmerman 1986). For example, we propose that investors’ tendency to evaluate earnings
relative to natural reference points (e.g., prior year earnings) may lead firms to manage earnings.
24
However, if managers’ compensation plans are structured such that they receive bonuses when
earnings exceed these same reference points, then managers’ efforts to manipulate earnings may
be due to contracting incentives predicted by economics. In focusing on the importance of
psychological, rather than economic, forces in this section, we are not suggesting that contracting
incentives or other economic forces are unimportant determinants of managers’ decisions to
manipulate earnings. Instead, our point is that investors’ preferences for particular earnings
patterns – preferences that some would consider economically irrational – may also play a
significant role. Consequently, we suggest that firms are likely to manage earnings even in the
absence of contracting incentives or other economic justifications. If this is true, then archival
researchers may obtain a more complete understanding of earnings management behaviors by
introducing such considerations into their models.
Understanding how disclosure format affects investors’ use of financial information
Accounting standards often afford managers flexibility in the format of their disclosures,
leading to cross-sectional variation in the placement, transparency, and labeling of some
financial disclosures. Archival research in financial accounting has shown that the format of an
accounting disclosure can influence whether and how investors use the information contained in
the disclosure. For example, Cotter and Zimmer (2003) show that investors are more likely to
positively value information about a firm’s asset revaluations when the information is recognized
in the financial statements rather than disclosed in the footnotes.
Economic theory offers an information signaling explanation for such phenomenon.
According to this signaling explanation, variations in the placement or description of information
may provide signals to decision makers about the meaning or significance of this information
(Spence 1973). For example, information contained in the footnotes is often less reliable than
25
information recognized on the financial statements (Hodge et al. 2003). Consequently,
management may attempt to signal that information is unreliable by choosing footnote disclosure
rather than financial statement recognition (Bernard and Shipper 1993). In other words,
investors may react differently to disclosures of different formats because there is information
content in the format itself. Existing archival research on format effects relies almost exclusively
on these types of information content arguments. For example, Cotter and Zimmer (2003) argue
that the reason investors place a higher premium on upward asset revaluations when they are
recognized in the balance sheet rather than disclosed in the footnotes is that they rationally infer
that recognized revaluations are more reliably measured than those that are merely disclosed.
Psychology theories, in contrast, suggest that format effects may arise even when a
disclosure’s format does not convey relevant information. Specifically, psychology offers two
additional explanations for format effects: 1) that format may influence investors’ decisions due
to differences in the relative ease of processing different formats, and 2) that certain formats may
lead to systematic biases in investors’ cognitive processing of the information disclosed. These
psychological explanations allow for new predictions about when and how format influences
investor reactions to the disclosures.
Most economists would not dispute that some disclosure formats are easier to process than
others. However, they would argue that if information contained in a firm’s disclosures is worth
processing (i.e., if money can be made from trading on that information), then the information
will be reflected in market prices. That is, even if a particular format is too difficult for most
investors to process, highly experienced expert investors will be willing and able to undo
superficial format differences in firms’ disclosures. Because these high-expertise investors are
26
assumed to set market prices, economic theory argues that superficial changes in information
format should not affect the extent to which the information is reflected in market prices.
Recent developments in psychology and behavioral finance question some of the
assumptions underlying these economic arguments. In particular, psychologists have shown that
people do not always properly assess costs and benefits when making cost/benefit trade-offs
(Kleinmuntz and Schkade 1993), suggesting that even high-expertise investors may sometimes
under-estimate the potential benefits of processing complex disclosures. Further, behavioral
finance studies suggest that less-expert “noise” traders are able to influence market prices and
that, in some situations, these traders are the primary driver of market prices (Schleifer 2000).
Thus, as long as some investors are unable to process format differences, informationally-
equivalent disclosures that vary only in their ease of processing (i.e., transparency) can have
differential effects on market prices (Payne et al. 1993).
Psychology studies also suggest that even highly expert and motivated investors are
sometimes vulnerable to superficial variations in information formats. We base this idea on
research, described in Section 2, revealing that the way information is presented systematically
affects the way people think about that information. Drawing on prospect theory, for example,
we predict that investors will react differently to financial information depending on the points of
comparison that explicitly accompany the financial data. That is, moderately positive financial
results may seem quite positive when presented alongside the relatively poor results of the
general industry, whereas those same results might seem disappointing when presented alongside
optimistic projections from the past. Other psychology research suggests that people are more
likely to recall and give weight to information that is located at the beginning or end of a
document (Hogarth and Einhorn 1992), displayed graphically (Miniard et al. 1991), and/or
27
repeated frequently (Unnava and Burnkrant 1991). Importantly, these types of format effects
often stem from hard-wired cognitive biases that have been shown to be unresponsive to
debiasing efforts (Arkes 1991). Consequently, novices and experts alike fall victim to such
information processing biases, even in tasks where there are strong incentives to get the right
answer (McNeil et al. 1982).
In sum, economic theories suggest one possible explanation for format effects – that the
format in which information is disclosed provides relevant information in and of itself.
However, psychology theories suggest other vehicles by which information format may
influence investors’ decisions. Specifically, psychologists show that format may affect decision-
making because some formats are more difficult to process and/or lead to systematically biased
processing. The idea that informationally-irrelevant variations in the presentation of financial
data can influence investor decision-making is significant for archival researchers because it
suggests that markets may be driven by disclosure features other than those that described by
economic models. Archival researchers can draw on the psychological literature on format
effects to identify and motivate new predictions about when and how various formats effects
(e.g., the location of information within a firm’s financial disclosures or the mode in which
information is presented) affect market prices.
Understanding when and how analysts’ incentives influence their research reports Much of the existing research on analyst expertise assumes that analysts have accuracy as
their goal. However, analysts also have incentives to please company managers, and these
incentives may influence earnings forecasts and stock recommendations (Francis and Philbrick
1993; McNichols and O’Brien 1997). Economic theories recognize the important role that
incentives play in behavior (Indjejikian 1999; Jensen and Meckling 1976). For example, if
28
analysts have incentives to maintain access to management, economic theories predict that
awareness of these incentives will drive analysts to engage in management-pleasing behaviors.
As explained in Section 2, research from social psychology predicts that there may be
additional paths by which incentives affect analysts’ forecasts and recommendations.
Specifically, this research suggests that incentives also can create subconscious biases in
analysts’ search for and processing of new information. Disentangling these conscious and
subconscious effects is important because it has implications for whether and how analysts’
reports can be debiased. For example, regulators’ recent attempts (e.g., Sarbanes-Oxley Act
2002) to curb analysts’ lack of independence with penalties for biased reporting will be
successful only to the extent that this non-independence is due to conscious actions on the part of
analysts.
One phenomenon that can create biased analysts’ judgments is subconscious biases in
analysts’ information search. Social psychologists have shown that when people search for new
information, they seek information that confirms their pre-existing beliefs. Importantly, this
“confirmation bias” exists even when people have incentives to be accurate (Arkes 1991). This
literature suggests that analysts’ incentives to please management will lead to biases in their
search for new information about the firm. Moreover, these effects will not be eliminated by
penalizing analysts’ for biased reporting, because analysts likely are not aware of the degree to
which management’s preferences are influencing their information search.
For example, analysts who have a preferred conclusion (e.g., a ‘Strong Buy’ stock
recommendation) will subconsciously seek information supporting this conclusion, even if they
are compensated based on their recommendation accuracy. In other words, analysts may ‘work
backward’ by looking for supporting analysis to justify a desired forecast or recommendation
29
rather than performing their analysis first and using the results to derive a forecast or
recommendation. Thus, while the accounting literature assumes that analysts ‘work forward’ in
providing earnings forecasts and stock recommendations (Bradshaw 2002), confirmatory search
processes imply analysts sometimes ‘work backward.’ This type of reasoning often was alleged
in the popular press during the bull market of the late 1990s. Analysts were accused of starting
with existing stock prices, and then providing supporting analysis (often with extreme
expectations for growth) that could support such valuations (Serwer 1998).
A related phenomenon that can create biases in analysts’ judgments is biased information
processing. As outlined in Section 2, social psychologists also have shown that people are more
critical of information that disconfirms a favored hypothesis than information that confirms the
hypothesis (Edwards and Smith 1996). This suggests that even when disconfirming (i.e.,
negative) information arises, analysts will discount that information. In other words, analysts
may “see” warning signals about future performance but discount this information until it
becomes incontrovertible. Similarly, analysts may begin to interpret ambiguous or negative
information as positive. For example, they may introduce valuation metrics or re-interpret
traditionally negative indicators in a positive way to justify their forecasts and recommendations.
Anecdotal evidence of such reasoning appeared during the 1990s dot.com craze. At that time,
some analysts introduced new valuation metrics such as revenue per subscriber and number of
page views, and re-interpreted traditionally negative factors (e.g., a high ‘cash burn’ rate) as
positive factors to justify high valuations for Internet stocks (Nocera and Maroney 1999;
Veverka 1999).
As noted in Section 2, biased search and processing can lead to belief perseverance and
polarization. Consequently, these phenomena also may help us understand why analysts who
30
receive identical financial data about a company sometimes reach completely different
conclusions about the company’s future earnings potential and stock performance. If two
analysts who have opposite starting beliefs about a company’s financial future receive a mixed
set of financial data about the company, the analysts may find those parts of the data that bolster
their prior beliefs to be particularly noteworthy and persuasive. The end result is that the mixed
financial evidence that the analysts receive may actually end up further polarizing the analysts’
beliefs about the company rather than bringing them closer together.12
Financial accounting examples where psychology theories allow for more specific predictions Understanding how managers develop reporting reputations
Managers are rightfully concerned about their reputations with analysts and investors (Healy
and Palepu 1995). Management’s reputation for credible disclosure affects its ability to convey
information to the capital markets (Williams 1996; Hirst et al. 1999), and more reputable firms
have lower information risk and thus a lower cost of capital (King et al. 1990; Botosan 1997;
Easley and O’Hara 2003). Given the importance of a company’s reporting reputation to its
financial well-being, it is critical to understand how these reputations form. Economic theories
describe the consequences of having (or not having) a positive reputation (Akerlof 1968; Spence
1973; Rothschild and Stiglitz 1976), but provide little guidance on the process by which
reputations form.13 Social psychologists have conducted hundreds of studies investigating when
and how people’s behaviors affect their credibility (e.g., Hovland et al. 1953; Mika 1981; Gilbert
1998). Consequently, these studies provide a rich set of theories that can be used to make
predictions about how managers’ disclosure decisions affect their reporting reputation.
A simplistic analysis of reputation formation might focus on the actions that a company takes
and the outcomes that it obtains. That is, companies that act in positive (negative) ways and that
31
obtain positive (negative) outcomes are relatively more likely to develop positive (negative)
reputations. However, social psychology’s correspondence inference theory suggests that
reputation formation is more complex – some actions taken by management are more likely to be
perceived as indicators of the management’s integrity than others. Specifically, correspondent
inference theory holds that people are more likely to assume that the actions of others reflect
dispositional forces when those actions appear to be voluntary and/or devoid of personal benefit
(Jones and Harris 1967). One implication of this theory is that management’s voluntary
disclosures will have greater effects on management’s reporting reputation than its required
disclosures. On the flip side, management may be able to manipulate its reporting reputation by
claiming that reports that appear to be misleading resulted from a lack of choice.
Another implication of correspondent inference theory is that forthcoming disclosures are
less likely to improve management’s reputation when the disclosures are perceived to be self-
serving. For example, if management’s pay is based on the firm’s financial performance,
management will personally benefit from disclosure of positive news. Correspondent inference
theory predicts that in such cases, there will be little reputational benefit to providing
forthcoming disclosures about positive news. However, timely reporting of negative news is
generally not self-serving, and so forthcoming negative news disclosures are predicted to
enhance management’s reporting reputation.14 In other words, correspondent inference theory
predicts that when investors assess management’s reporting reputation, they will attach greater
weight to management’s negative news reporting practices than its positive news reporting
practices.
Other areas within psychology also provide guidance on how reputations form. For example,
belief-updating studies predict that investors will use confirmatory strategies when faced with
32
new evidence about management’s reporting reputation (cf. Snyder and Swann 1978; Tesser
1978). As described in Section 2, psychologists have shown that people tend to search for and
place more weight on evidence that confirms, rather than tests, an initial hypothesis (Klayman
and Ha 1987). This suggests that the earliest information that investors and analysts receive
about management will provide a lens through which all subsequent information is filtered.
Evidence that confirms investors’ initial beliefs about management will easily find its way
through the lens and reinforce initial beliefs. Evidence that disconfirms investors’ beliefs will
not receive the same attention and respect.
Consider a simple example. If investors and analysts believe that company management is
forthcoming, then they may be more inclined to provide a management-favorable interpretation
of subsequently received evidence that challenges this view. They may question the importance
of an omission in the company’s financial statements, or suggest innocent interpretations of the
omission (e.g., unintentional recording error, management was unaware of the omitted event at
the time the report was written). In contrast, if investors and analysts believe that the company’s
management is not entirely forthcoming, they are likely to be less charitable in their
interpretations of the omission, treating it as proof of their suspicions. Moreover, these early
doubters may look for other reasons and evidence that bolsters their initial skepticism. One
result of such confirmatory search strategies is management’s early disclosure decisions are
especially important, because investors’ initial impressions about management will be quite
resistant to change (Anderson 1980; Darley and Gross 1983).
To summarize, although economic theories provide useful guidance on the consequences of
having or not having a positive reputation, they provide little guidance on how managers’
specific disclosure decisions affect their reputation for credible disclosure. Psychology theories
33
may help fill this void, as there is a large body of descriptive evidence within psychology on how
a person’s actions affect his or her reputation. We described how correspondent inference theory
and belief updating theories from social psychology allow for a number of specific predictions
about which management disclosures will have the greatest effects on management’s reporting
reputation. Correspondence inference theory suggests, for example, that management’s
voluntary disclosures about negative news will be key determinant of management’s reporting
reputation, and studies on belief updating outline the importance of management’s early
disclosure decisions.
Understanding investor reactions to changes in disclosure policy
Economic theory suggests that when firms commit to increased levels of disclosure, this
commitment will reduce information asymmetry and, in turn, the firm’s cost of capital (Diamond
and Verrecchia 1991; Baiman and Verrecchia 1996). However, archival studies that investigate
these relations show mixed results (see Leuz and Verrecchia 2000 for a review). One potential
reason for these mixed results is that investors must believe that management is making a
“commitment” to increased disclosure for there to be effects on the cost of capital (Leuz and
Verrecchia 2000, 92). Archival studies have typically assumed that investors will perceive any
increase in a firm’s disclosure level to be a commitment to greater disclosure.15 However, it
seems likely that investors do not perceive all additional disclosure as a commitment. Therefore,
tests of the effects of expanded disclosure could be improved with a better understanding of
when investors interpret more disclosure as a commitment to increased levels of disclosure.
As outlined in Section 2, covariation theory predicts that three features of a disclosure – its
consensus, consistency, and distinctiveness – will affect investors’ beliefs about why the
disclosure was made. Applying this theory to the financial disclosure domain suggests that
34
investors are likely to evaluate disclosure changes, in part, based on how other companies
disclose similar events (disclosure consensus), whether the company continues to disclose this
type of information (disclosure consistency), and the company’s other disclosures (disclosure
distinctiveness). Investors’ beliefs about disclosure consensus, consistency, and distinctiveness,
in turn, will affect their perceptions about the likelihood that the expanded disclosure will persist
in future periods. Consequently, covariation theory should be useful for predicting how
investors will react when a company changes its disclosure policy.
To illustrate, consider the recent changes in General Electric’s disclosures. In the wake of
the Enron accounting scandal, General Electric began to provide a number of additional
disclosures in its annual report. The company stated that these new disclosures were part of an
effort to increase disclosure transparency (Silverman 2002). Covariation theory predicts that
investors’ reactions to this change will depend, in part, on the perceived consensus, consistency,
and distinctiveness of the increased disclosure.
For example, unfortunately for General Electric, other companies (e.g., IBM and Krispy
Kreme Doughnuts) increased their disclosures at about the same time. According to covariation
theory, greater disclosure by these other companies increases the consensus associated with the
behavior. High consensus is ordinarily more conducive to external rather than internal
attributions for a behavior. Therefore, investors are less likely to attribute General Electric’s
increased disclosure to an increased commitment to transparent disclosure (as management had
hoped) and more likely to attribute the behavior to environmental pressures. Such an attribution
may result in skepticism among investors about whether General Electric will continue its
expanded disclosures once the public scrutiny fades.
35
Although companies like General Electric cannot control consensus information (i.e., they
cannot control what other companies disclose), they can promote internal attributions by
reducing the distinctiveness of the disclosure change. For example, if General Electric
simultaneously improves its disclosure transparency in other venues such as earnings forecasts
and conference calls, investors will be more willing to believe that the additional annual report
disclosures reflect a commitment to more forthcoming disclosure. Thus, covariation theory
predicts that researchers are more likely to observe changes in information asymmetry and cost
of capital when disclosure changes are reflected in more than one area of their disclosures.
Covariation theory also suggests that changes in a firm’s disclosure level may not be
reflected immediately in the firm’s cost of capital, because it takes time for investors to
determine whether disclosure change is consistent. That is, investors will be better able to
evaluate whether disclosure increases reflect a true commitment to greater disclosure once they
know whether the disclosure changes persist in future periods. For example, while General
Electric may experience some immediate benefit from their increased disclosure, it is more likely
that any benefits will occur slowly over time, as investors become more convinced that the
disclosure will continue.
In sum, covariation theory from social psychology allows for specific predictions about when
changes in a firm’s disclosures will be reflected in the firm’s cost of capital. It suggests that
investors are more likely to infer a commitment to greater disclosure 1) when other firms are not
simultaneously increasing their disclosure levels (i.e., when there is greater perceived disclosure
distinctiveness), 2) when the firm increases its disclosure level in more than one area (i.e., when
there is greater perceived disclosure consensus), and 3) over time, as investors become more
36
confident that the increased disclosures will persist (i.e., when there is greater perceived
disclosure consistency).
Understanding how financial analysts develop expertise
Analysts are important information intermediaries in financial markets. They analyze the
financial data provided by firms and provide earnings forecasts and stock recommendations
based on their analyses. The financial press suggests that some analysts perform better than
others in these tasks (e.g., Institutional Investor provides a list of “All American” analysts and
The Wall Street Journal provides a list of “All Star” analysts), and the academic literature
confirms that analysts differ in expertise (Stickel 1992; Sinha et al. 1997).
Several archival papers have investigated why some analysts perform better than others
(Mikhail et al. 1997; Clement 1999; Jacob et al. 1999). These papers provide somewhat
conflicting results. Mikhail et al. (1997) offers a ‘learning by doing’ model that predicts that
analysts with more experience will issue more accurate forecasts. This model suggests that as
people practice a task, their performance improves (e.g., Anderson 1987; Seifert et al. 1997). In
support, Mikhail et al. examine analysts’ forecasts errors over time and report that an analyst’s
errors become smaller as he or she gains experience. Clement (1999) relies on similar theoretical
arguments and also finds that more experienced analysts have smaller forecast errors. However,
Jacob et al. (1999) suggest that the association between experience and forecast accuracy results
from survival bias. That is, more experienced analysts perform better not because they learn
from experience, but because weaker performing analysts are forced out of the profession. In
support, Jacob et al. find that when they include a dummy variable for each individual analyst in
their regressions, the effect of experience on analyst forecast errors becomes insignificant.
37
Psychology-based theories of expertise provide insight into these seemingly contradictory
findings and suggest new directions for research in this area.16 As noted in Section 2,
psychologists have shown that experience enhances expertise via its effects on knowledge
(Hunter 1983; Schmidt et al. 1986; see also Libby and Luft 1993). In other words, experience
does not directly affect performance. Rather, experience provides an opportunity for knowledge
acquisition, and greater knowledge improves performance. Psychology theories also highlight
the importance of innate ability in task performance (Sternberg 1997). Specifically, psychology
theories predict that in addition to having a direct effect on performance, ability will indirectly
affect performance because greater ability allows for more knowledge acquisition.
As previously noted, Jacob et al. (1999) show that when individual dummy variables for each
analyst are included in their regressions, the association between experience and performance
disappears. They conclude that experience does not affect financial analyst performance.
However, if these individual analyst dummy variables are capturing differing levels of know-
ledge between analysts, it is possible that experience still may be important in explaining analyst
performance (via its effects on analyst knowledge). Expertise theories suggest that more
complex regressions models – such as path analyses that enable the researcher to simultaneously
test a series of causal links – may be necessary when examining how analyst characteristics such
as experience and ability affect performance.
Another point archivalists may wish to consider is that person-specific predictors of expertise
such as knowledge and ability tend to be indicators of performance in complex (rather than
simple) tasks. The earnings forecasts and stock recommendations that analysts provide would
seem to be complex due to the large number of decision inputs and the lack of specified
procedures for how the inputs should be combined to arrive at a final judgment (Feltovich et al.
38
1984; Bonner 1994). We agree with this general point, but note that there is significant variation
in the complexity of analysts’ predictions across firms and across time. For example, some firms
have less-variable financial results and/or provide specific guidance to analysts. In these cases,
the analysts’ job is less complex and we might not expect to observe large performance
differences as a function of analyst expertise. Complexity also will vary over time because
financial results are more variable during volatile economic times. Finally, psychology research
suggests that high expertise analysts may underperform low expertise analysts under certain
conditions. Frensch and Sternberg (1989) called attention to this possibility when they argued
that experts who have well-developed problem-solving routines may be less flexible in their
methodologies and ways of thinking than those who are less experienced. This lack of flexibility
may cause highly experienced financial analysts to perform worse than less-experienced analysts
in unusual forecasting environments (Feltovich et al. 1997).
To summarize, expertise theories from psychology allow for more specific predictions about
the determinants and consequences of analyst expertise. Specifically, these theories suggest that
the relations between experience, ability, and expertise are more complex than assumed in much
of the existing financial accounting literature. Expertise theories also identify conditions where
analyst expertise is most likely to translate into superior performance. Specifically, they suggest
that expert analysts will tend to perform better than non-expert analysts and that these
performance differences will be greatest for difficult tasks. However, they also show that these
effects may reverse under highly unusual task conditions, where expert analysts’ well-developed
problem-solving techniques can lead to worse performance than that of less experienced (but
more flexible) non-experts.
39
Concluding Remarks Researchers from both psychology and economics seek to understand and describe human
behavior. However, as one well-known psychologist put it, psychology and economics often
seem to be “two disciplines divided by a common interest” (D. Gilbert, personal communication,
July 20, 2000). This divide also carries over into applied domains such as accounting. We find
that archival accounting researchers rely almost exclusively on economics-based theories,
despite the potential usefulness of psychology theories for their research.
We argue that this dearth of psychology-based archival accounting research is not necessarily
due to archival researchers’ lack of interest in psychological explanations, but rather a lack of
familiarity with psychology theories and how these theories can be productively applied. Our
paper bridges this gap by describing a number of relevant psychology theories and providing
examples of how these theories can be used to make different and/or more specific predictions
on financial accounting issues. Clearly, one paper cannot cover all of the psychology theories
that might be useful to accounting researchers. However, we take an important step in
familiarizing archival researchers with the types of issues studied by psychologists and provide
illustrations of how these ideas can be applied in the financial accounting domain.
Although psychology offers a wealth of insights for archival researchers, there are some
limitations on its usefulness. First, psychology theories are not useful in all areas of archival
accounting research. That is, because psychologists focus on understanding human behavior,
accounting issues that do not deal with human behavior are unlikely to be aided by psychology
theories. For example, psychology theories are less relevant to researchers who study the
statistical properties of earnings than to researchers who study how investors respond to a
particular earnings number. It is important to reiterate that we are not advocating that
40
psychology theories replace economic theories in financial accounting research. Rather, we
believe that economic and psychology theories should be treated as complementary, and the most
suitable theory or theories be used.
Second, some areas within psychology are still developing and thus cannot be used to
generate definitive predictions about behaviors in the accounting domain. For example,
psychologists have identified that people sometimes over-react to new information (Koehler
1996) and sometimes under-react to new information (Phillips and Edwards 1966; Lord et al.
1979). This literature should prove useful to archival accounting researchers who study over-
and under-reaction to accounting information (e.g., Bernard and Thomas 1989). However,
because psychology researchers are still trying to understand the conditions under which people
over- and under-react, this literature may not be immediately helpful for generating precise
predictions about when investors and analysts will over-react and/or under-react to accounting
information.
A third potential limitation (but also a potential opportunity) is that psychology theories tend
to describe individual-level behaviors, but do not specify how these behaviors translate into
aggregate market prices. As noted in the introduction to the paper, archival researchers are in a
unique position to incorporate psychology theories into their existing knowledge of market
reactions to accounting information to advance our understanding of this issue. Although
psychology theories tend to describe individual-level behaviors, these behaviors are systematic
behavioral tendencies rather than idiosyncratic behaviors, and in many cases, these systematic
tendencies translate directly into aggregate behavior (Tuttle et al. 1997). Understanding when
the individual-level behaviors observed by psychologists translate into market-level behavior and
when they do not appears to be an important area for future research (Lee 2001).
41
Finally, archival researchers may wonder about the feasibility of testing psychology-based
predictions as compared to economics-based predictions. Testing psychology-based predictions
may be hindered in some cases by the availability of archival data. For example, psychology
theories about the content of human memory are not likely candidates for archival tests unless
those theories address an observable implication of that memory organization. In other cases,
only rough proxies may be available for the construct of interest. For example, it is difficult to
find archival data measuring a financial analyst’s innate ability (Clement 1999). However, this
problem is not exclusive to psychology-based archival research – archival researchers are often
required to find creative proxies for construct-level variables, even when testing economics-
based predictions. Not surprisingly, experimental researchers also face similar data availability
issues. For example, experimentalists cannot directly measure an individual’s innate ability and
so must contend with problems of construct validity. We conjecture that as the use of
psychology theory grows, archival researchers will develop and refine proxies for psychological
constructs, as they have done so over the years with economic constructs.
In sum, we acknowledge that introducing a new theoretical framework to archival accounting
research is not a cost-free endeavor. However, we believe that it will be well worth the effort, as
greater use of psychology theories will result in more descriptive models of the financial
reporting process.
42
ENDNOTES 1 When we use the term ‘archival,’ we are referring to research that tests hypotheses using historical data such as stock prices, management disclosures, and analyst forecasts. Although we limit our focus in this paper to archival research, many of our comments also could apply to analytical modeling research. 2 A survey of Ph.D. student participants at the 2002 American Accounting Association doctoral consortium found that while virtually all Ph.D. students had taken or expected to take a financial archival seminar, only half had taken or expected to take a seminar on behavioral accounting topics (Kinney 2003). Presumably, even fewer take doctoral-level psychology courses. 3 Some progress has been made on this issue; a small number of archival researchers now have psychology training and/or exposure to psychology research from experimentalist colleagues, and this exposure is reflected in their research (e.g., Hand 1990; Mikhail et al. 1997; Plumlee 2003). However, the vast majority of archival researchers still have little exposure to psychology research. This unfamiliarity with the psychology literature is problematic because makes it difficult for archival researchers to recognize where psychology theories can be productively applied. 4 For those interested in a thorough review of the psychology-based experiments in the applied domain of financial accounting, see a recent review paper by Libby et al. (2001). 5 In fact, Daniel Kahneman was recently awarded the Nobel Prize in Economics for his research in this area. 6 In 1954, Leonard Savage demonstrated that the axioms that give expected utility framework its normative status are met even when a decision maker’s subjective estimate of p, denoted s(p), is substituted for p. Savage’s variant of EU is called subjective expected utility (SEU). 7 Interestingly, studies show that even experts in optimal feedback conditions show poor performance when they are assessing very low base rate events (e.g., earthquakes) because there are few data to challenge incorrect predictions (Lichtenstein et al. 1982). 8 More surprising, perhaps, is evidence that those who have the most expertise are often no more proficient in their activities or accurate in their forecasts than those who have some, but less, expertise. This prompted one researcher to offer the following rule of thumb for those in search of expertise: hire the cheapest expert you can find (Armstrong 1980). 9 McKinney (1991) offered a dramatic illustration of this point in his study of how novice and expert U.S. Air Force pilots reacted to actual emergencies and crises on their planes. Emergencies refer to serious problems that pilots have seen in their training (e.g., engine shut down) and crises refer to serious problems that are so unusual that the pilots have not encountered them in their training (e.g., wing falls off). McKinney found that although experts outperformed novices in emergencies, novices were quicker than experts to recognize crises for what they were and bail out of failing airplanes. Tragically, many expert pilots mistook crises for solvable emergencies and failed to recognize the novelty and severity of the problems until it was too late. Thus, under certain conditions, greater expertise may actually be detrimental to performance. 10 These differences may be due, in part, to the different ways that economic and psychology theories are derived. Many economic theories are derived deductively. That is, economists tend to begin with a set of assumptions about how people should behave and develop general models of behavior based on these assumptions. In contrast, psychology theories are generally derived inductively from descriptive evidence
43
on people’s actual behaviors. Because of this, psychology theories do not necessarily assume that behaviors are the product of rational forces. 11 Specifically, prospect theory suggests that investors’ preferences for particular patterns of earnings will encourage firms to: 1) avoid reporting small losses, 2) take the occasional “big bath,” 3) create “cookie jar reserves” during very strong years, and 4) attempt to influence the reference points that investors and analysts use to assess reported earnings. 12 This phenomenon appears to contradict a commonsense normative standard of belief updating. That is, if two analysts receive identical information, their views should converge rather than diverge. However, this simplistic normative view should be tempered by an awareness that there may not be clear standards for evaluating the importance (or weight) that should be assigned to new information about the firm. Some have argued that the Bayesian belief updating actually requires that decision makers draw upon their prior beliefs when trying to determine this weighting (Koehler 1993), because in general, it is reasonable for decision makers to assume that information that agrees with a prior belief is more reliable than information that does not. Carrying this logic through, it may be that belief polarization among analysts that results from different weightings of the mix of financial information they receive is not only psychological understandable but economically appropriate as well. 13 Some economic studies identify the conditions necessary for reputations to form (Akerlof 1968; Rotschild and Stiglitz 1976; Spence 1973). However, to the best of our knowledge, few, if any, economic theories detail how specific actions or inactions affect reputation. 14 Managers occasionally have greater incentives to provide negative news, such as when they are trying to lower the strike price of their employee stock options (Aboody and Kasznik 2000). 15 Archival studies often measure a firm’s commitment to increased disclosure based on its overall quantity of disclosure. 16 Isolating the relative roles of experience and natural ability/aptitude on financial analyst performance is important because it has implications for whether training can significantly improve analysts’ skills.
44
REFERENCES
Aboody, D. and R. Kasznik. 2000. CEO stock option awards and the timing of corporate voluntary disclosures. Journal of Accounting and Economics 29 (1): 73-100.
Akerlof, G. 1968. The market for "lemons": Quality uncertainty and the market mechanism.
Quarterly Journal of Economics 84 (3): 488-500. Allais, M. 1953. Le comportement de l’homme rationnel devant le risqué, critique des postulats
et axioms de l’ école Américaine. Econometrica 21 (4): 503-546. Allport, G. 1954. The nature of prejudice. New York: Doubleday Books. Anderson, J. 1980. Cognitive Psychology and Its Implications. San Francisco: Freeman. Anderson, J. 1987. Skill acquisition: Compilation of weak-method problem solutions.
Psychological Review 94 (2): 192-210. Arkes, H. 1991. Costs and benefits of judgment errors: Implications for debiasing. Psychological
Bulletin 110 (3): 486-498. Armstrong, S. 1980. The seer-sucker theory: The value of experts in forecasting. Technology
Review 83: 18-24. Arrow, K. 1982. Risk perception in psychology and economics. Economic Inquiry 20 (1): 1-9. Baiman, S., and R. Verrecchia. 1996. The relation among capital markets, financial disclosure,
production efficiency, and insider trading. Journal of Accounting Research 34 (1): 1-122. Balla, J., R. Iansek, and A. Elstein. 1985. Bayesian diagnosis in presence of pre-existing disease.
The Lancet 1 (8424): 326-329. Bernard, V., and K. Schipper. 1993. Recognition and Disclosure in Financial Reporting.
Working paper, University of Chicago. Bernard, V., and J. Thomas. 1989. Post-earnings announcement drift: Delayed price response or
risk premium? Journal of Accounting Research 27 (Supplement): 1-36. Birnbaum, M., and S. Stegner. 1979. Source credibility in social judgment: Bias, expertise, and
the judge's point of view. Journal of Personality and Social Psychology 37 (1): 48-74. Bonner, S. 1994. A model of the effects of audit task complexity. Accounting, Organizations,
and Society 19 (3): 213-234. Botosan, C. 1997. Disclosure level and the cost of equity capital. The Accounting Review 72 (3):
323-349.
45
Bradshaw, M. 2002. The use of target prices to justify sell-side analysts’ stock recommendations. Accounting Horizons 16 (1): 27-41.
Burgstahler, D. and I. Dichev. 1997. Earnings management to avoid earnings decreases and
losses. Journal of Accounting and Economics 24 (1): 99-126. Cadsby, T. 2000. The Ten Biggest Investment Mistakes Canadians Make and How to Avoid
Them. Toronto: Stoddart. Camerer, C. 1987. Do biases in probability judgments matter in markets? Experimental
evidence. The American Economic Review 77 (5): 981-997. Camerer, C., L. Babcock, G. Loewenstein, and R. Thaler. 1997. Labor supply of New York City
cabdrivers: One day at a time. Quarterly Journal of Economics 112 (2): 407-441.
Camerer, C. and R. Hogarth. 1999. The effects of financial incentives in experiments: A review and capital-labor-production framework. Journal of Risk and Uncertainty 19 (1-3): 7-42.
Chase, W., and H. Simon. 1973. The minds’ eye in chess. In Visual Information Processing, ed.
W. Chase, Chapter 5. New York: Academic Press. Chi, M., P. Feltovich, and R. Glazer. 1981. Categorization and representation of physics
problems by experts and novices. Cognitive Science 5 (2): 121-152. Chi, M., R. Glaser, and E. Rees. 1982. Expertise in problem solving. In Advances in the
Psychology of Human Intelligence, ed. R. Sternberg, 7-76. Hillsdale, NJ: Erlbaum. Clement, M. 1999. Analyst forecast accuracy: Do ability, resources, and portfolio complexity
matter? Journal of Accounting and Economics 27 (3): 285-303. Cotter, J., and Zimmer, I. 2003. Disclosure versus recognition: The case of asset revaluations.
Asia-Pacific Journal of Accounting and Economics, forthcoming. Daniel, K., D. Hirshleifer, and S. Teoh. 2002. Investor psychology in capital markets: Evidence
and policy implications. Journal of Monetary Economics 49 (1): 139-210. Darley, J., and P. Gross. 1983. A hypothesis confirming bias in labeling effects. Journal of
Personality and Social Psychology 44: 20-33. Defond, M. and J. Jiambalvo. 1994. Debt covenant effects and the manipulation of accruals.
Journal of Accounting and Economics 17 (1): 145-176. Degeorge, F., J. Patel, and R. Zeckhauser. 1999. Earnings management to exceed thresholds.
Journal of Business 72 (1): 1-33. Diamond, D., and R. Verrecchia. 1991. Disclosure, Liquidity, and the Cost of Capital. Journal
of Finance 46 (4): 1325-1359.
46
Doherty, M, C. Mynatt, R. Tweney, and M. Schiavo. 1979. Pseudodiagnosticity. Acta
Psychologica 43 (2): 111-121. Dowie, J. 1976. On the efficiency and equity of betting markets. Economica 43: 139-150. Eagly, A., W. Wood and S. Chaiken. 1978. Causal inferences about communicators and their
effect on opinion change. Journal of Personality and Social Psychology 36 (4): 424-435. Easley, D. and M. O’Hara. 2003. Information and the cost of capital. Forthcoming in Journal of
Finance. Eddy, D. 1982. Probabilistic reasoning in clinical medicine: Problems and opportunities. In
Judgment under Uncertainty: Heuristics and Biases, eds. D. Kahneman, P. Slovic, and A. Tversky, 249-267. Cambridge, UK: Cambridge University Press.
Edwards, K. and E. Smith. 1996. A disconfirmation bias in the evaluation of arguments. Journal
of Personality and Social Psychology 71 (1): 5-24. Einhorn, H. 1980. Learning from experience and suboptimal rules in decision making. In
Cognitive Processes in Choice and Decision Behavior, ed. T. Wallsten, 1-20. Lawrence Erlbaum Associates: Hillsdale, NJ.
Einhorn, H. and R. Hogarth. 1978. Confidence in judgment: Persistence of the illusion of
validity. Psychological Review 85 (5): 395-416. Ellsberg, D. 1961. Risk, ambiguity, and the Savage axioms. Quarterly Journal of Economics 75
(4): 643-669. Feltovich, P., P. Johnson, J. Moller, and D. Swanson. 1984. LCS: The role and development of
medical knowledge in diagnostic expertise. In Readings in medical artificial intelligence: The first decade, eds. W. Clancey and E. Shortliffe, 275-319. Reading, MA: Addison Wesley.
Feltovich, R., K. Ford, and R. Hoffman (eds.) 1997. Expertise in Context. Menlo Park, CA: The
MIT Press. Feltovich, P., R. Spiro, and R. Coulson. 1997. Issues of expert flexibility in contexts
characterized by complexity and change. In Expertise in Context, eds. R. Feltovich, K. Ford, and R. Hoffman, 125-146. Menlo Park, CA: The MIT Press.
Festinger, L. 1957. A Theory of Cognitive Dissonance. Palo Alto, CA: Stanford University
Press. Fields, T., T. Lys, and L. Vincent. 2001. Empirical research on accounting choice. Journal of
Accounting and Economics 31 (1-3): 255-307.
47
Francis, J. and D. Philbrick. 1993. Analysts’ decisions as products of a multi-task environment. Journal of Accounting Research 31 (2): 216-230.
Frensch, P. and R. Sternberg. 1989. Expertise and intelligent thinking: When is it worse to know
better? In Advances in the Psychology of Human Intelligence, ed. R. Sternberg, 157-188. Hillsdale, NJ: Erlbaum.
Gilbert, D. 1998. Ordinary personology. In The Handbook of Social Psychology, eds. D.
Gilbert, S. Fiske, and G. Lindzey, 89-150. Boston, MA: McGraw-Hill. Glare, P., K. Virik, M. Jones, M. Hudson, S. Eychmuller, J. Simes, and N. Christakis. 2003. A
systematic review of physicians’ survival predictions in terminally ill cancer patients. British Medical Journal 327 (7408): 195-197.
Glaser, R. 1990. The reemergence of learning-theory within instructional research. American
Psychologist 45 (1): 29-39. Hand, J. 1990. A test of the extended functional fixation hypothesis. The Accounting Review 65
(4): 740-763. Healy, P. 1985. The effect of bonus schemes on accounting decisions. Journal of Accounting
and Economics 7 (1-3): 85-107. Healy, P. and K. Palepu. 1995. The challenge of investor communication: The case of CUC
International. Journal of Financial Economics 38 (2): 111-140. Healy, P. and J. Wahlen. 1999. A review of the earnings management literature and its
implications for standard setting. Accounting Horizons 13 (4): 365-383. Heider, F. 1958. The Psychology of Interpersonal Relations. New York: Wiley. Hirst, E., L. Koonce, and J. Miller. 1999. The joint effect of management’s prior forecast
accuracy and the form of its financial forecasts on investor judgment. Journal of Accounting Research 37 (Supplement): 101-124.
Hodge, F., J. Kennedy, and L. Maines. 2003. Does search-facilitating technology improve the
transparency of financial reporting? Working paper, University of Washington. Hogarth, R., and H. Einhorn. 1992. Order effects in belief updating: The belief-adjustment
model. Cognitive Psychology 24 (1): 1-55. Hogarth, R., and S. Makridakis. 1981. Beyond discrete biases: Functional and dysfunctional
aspects of judgmental heuristics. Psychological Bulletin 90 (2): 197-217. Hovland, C., I. Janis, and H. Kelley. 1953. Communication and Persuasion. Yale University
Press: New Haven, Connecticut.
48
Hunter, J. 1983. A causal analysis of cognitive ability, job knowledge, job performance, and supervisory ratings. In Performance Measurement and Theory, eds. F. Landy, S. Zeddeck, and J. Cleveland, 257-266. Hillsdale, NJ: Erlbaum.
Indjejikian, R. 1999. Performance evaluation and compensation research: An agency
perspective. Accounting Horizons 13 (2): 147-157. Jacob, J., T. Lys, and M. Neale. 1999. Expertise in forecasting performance of security analysts.
Journal of Accounting and Economics 28 (1): 51-82. Jensen, M., and W. Meckling. 1976. Theory of the firm: Managerial behavior, agency costs and
ownership structure. Journal of Financial Economics 3 (4): 305-360. Jones, E. and K. Davis. 1965. From acts to disposition: The attribution process in person
perception. In Advances in Experimental Social Psychology: Volume 2, ed. L. Berkowitz, 219-266. Academic Press: New York.
Jones, E., and V. Harris. 1967. The attribution of attitudes. Journal of Experimental Social
Psychology 3: 1-24. Jones, E. and R. Nisbett. 1972. The actor and the observer: Divergent perceptions of the causes
of behavior. In Attribution: Perceiving the Causes of Behavior, eds. E. Jones, D. Kanouse, H. Kelley, R. Nisbett, S. Valins, and B. Weiner, 79-94. Morristown, NJ: General Learning Press.
Joyce, E. and G. Biddle. 1981. Are auditors’ judgments sufficiently regressive? Journal of
Accounting Research 19 (2): 323-349. Kachelmeier, S., and M. Shehata. 1992. Examining risk preferences under high monetary
incentives: Experimental evidence from the Peoples-Republic-of-China. American Economic Review 82 (5): 1120-1141.
Kahneman, D., and A. Tversky. 1979. Prospect theory: An analysis of decision under risk.
Econometrica 47 (2): 263-291. Kelley, H. 1967. Attribution theory in social psychology. In Nebraska Symposium on
Motivation, ed. D. Levine, 192-241. Lincoln, Nebraska: University of Nebraska Press. Kelley, H. 1972. Causal schemata and the attribution process. In Attribution: Perceiving the
Causes of Behavior, eds. E. Jones, D. Kanouse, H. Kelley, R. Nisbett, S. Valins, and B. Weiner, 151-174. Morristown, NJ: General Learning.
Key, K. 1997. Political cost incentives for earning management in the cable television industry.
Journal of Accounting and Economics 23 (3): 309-337. Kida, T. 1984. The effect of causality and specificity on data use. Journal of Accounting
Research 22 (1): 145-152.
49
King, R., G. Pownall, and G. Waymire. 1990. Expectations Adjustment via Timely Management Forecasts: Review, Synthesis, and Suggestions for Future Research. Journal of Accounting Literature: 113-44.
Kinney, W. 2003. “New accounting scholars—Does it matter what we teach them?” Issues in
Accounting Education 18 (1): 37-47. Klayman, J. and Y. Ha. 1987. Confirmation, disconfirmation, and information in hypothesis
testing. Psychological Review 94 (2): 211-228. Kleinmuntz, D., and D. Schkade. 1993. Information displays and decision processes.
Psychological Science 4 (4): 221-227. Koehler, D., L. Brenner, and D. Griffin. 2002. The calibration of expert judgment: Heuristics and
biases beyond the laboratory. In T. Gilovich, D. Griffin, and D. Kahneman (eds.) (pp. 686-715). Heuristics and Biases: The Psychology of Intuitive Judgment. Cambridge: Cambridge University Press.
Koehler, J. 1993. The influence of prior beliefs on scientific judgments of evidence quality.
Organizational Behavior and Human Decision Processes 56: 28-55. Koehler, J. 1996. The base rate fallacy reconsidered: Descriptive, normative, and
methodological challenges. Behavioral and Brain Sciences 19 (1): 1-53. Koriat, A, S. Lichtenstein, and B. Fischhoff. 1980. Reasons for confidence. Journal of
Experimental Psychology: Human Learning and Memory 6 (2): 107-118. Kothari, S.P. 2001. Capital markets research in accounting. Journal of Accounting and
Economics 31 (1-3): 105-231. Lee, C. 2001. Market efficiency and accounting research. A discussion of ‘capital market
research in accounting’ by S.P. Kothari. Journal of Accounting and Economics 31 (1-3): 233-253.
Leuz, C. and R. Verrecchia. 2000. The economic consequences of increased disclosure. Journal
of Accounting Research 38 (1): 91-124. Libby, R. and J. Luft. 1993. Determinants of judgment performance in accounting settings:
Ability, knowledge, motivation, and environment. Accounting, Organizations and Society 18 (5): 425-450.
Libby, R., R. Bloomfield, and M. Nelson. 2002. Experimental research in financial accounting.
Accounting, Organizations and Society 27 (8): 777-812. Lichtenstein, S., B. Fischhoff, and L. Phillips. 1982. Calibration of probabilities: The state of the
art to 1980. In Judgment Under Uncertainty: Heuristics and Biases, eds. D. Kahneman, P. Slovic, and A. Tversky, 306-334. Cambridge, UK: Cambridge University Press.
50
Lichtenstein, S., and P. Slovic. 1971. Reversals of preference between bids and choices in
gambling decisions. Journal of Experimental Psychology 89 (1): 46-55. Lord, C., M. Lepper, and D. Mackie. 1984. Attitude prototypes as determinants of attitude-
behavior consistency. Journal of Personality and Social Psychology 46 (6): 1254-1266. Lord, C., L. Ross, and M. Lepper. 1979. Biased assimilation and attitude polarization: The
effects of prior theories on subsequently considered evidence. Journal of Personality and Social Psychology 37 (11): 2098-2109.
Markowitz, H. 1952. The utility of wealth. The Journal of Political Economy, 60 (2): 151-158. Matsumoto, D. 2002. Management’s incentives to avoid negative earnings surprises. The
Accounting Review 77 (3): 483-514. McKinney, E. 1991. Experience, Uniqueness of Task, and Decision Making Performance : A
Retrospective Study of Air Force Pilots and Inflight Aircraft Malfunctions, 1980-1990. Dissertation Abstracts International (UMI No. ATT 9212591).
McNeil, B., S. Pauker, H. Sox, and A. Tversky. 1982. On the elicitation of preferences for
alternative therapies. New England Journal of Medicine 306: 1259-1262. McNichols, M. and P. O’Brien. 1997. Self-selection and analyst coverage. Journal of
Accounting Research 35 (Supplement): 167-208. Mika, S. 1981. Some determinants of source credibility. Polish Psychological Bulletin 12(2):
79-86. Mikhail, M., B. Walther, and R. Willis. 1997. Do security analysts improve their performance
with experience? Journal of Accounting Research 35 (Supplement): 131-157. Miniard, P., S. Bhatla, K. Lord, P. Dickson, and H. Unnava. 1991. Picture-based persuasion
processes and the moderating role of involvement. Journal of Consumer Research 18 (1): 92-107.
Murphy, A., and R. Winkler. 1984. Probability forecasting in meteorology. Journal of the
American Statistical Association 79 (387): 391-400. Neale, M., and G. Northcraft. 1986. Experts, amateurs, and refrigerators: Comparing expert and
amateur negotiators in a novel task. Organizational Behavior and Human Decision Processes 38 (3): 305-317.
Nocera, J. and T. Maroney. 1999. Do you believe? How Yahoo! became a blue chip: A tale of
how Wall Street and the rest of us learned to stop worrying and love an insanely valued Internet stock. Fortune (June 7): 76.
51
Payne, J., J. Bettman, and E. Johnson. 1993. The Adaptive Decision Maker. New York: Cambridge University Press.
Phillips, L. and W. Edwards. 1966. Conservatism in a simple probability inference task. Journal
of Experimental Psychology 72: 346-354. Plumlee, M. 2003. The effect of information complexity on analysts’ use of that information.
The Accounting Review 78 (1): 275-296. Rabin, M. 1998. Psychology and economics. Journal of Economic Literature 36 (1): 11-46. Redelmeier, D., and E. Shafir. 1995. Medical decision-making in situations that offer multiple
alternatives. Journal of the American Medication Association 273 (4): 302-305. Ross, L. 1977. The intuitive psychologist and his shortcomings: Distortions in the attribution
process. In Advances in Experimental Social Psychology: Volume 10, ed. L. Berkowitz, 174-221. New York: Academic Press.
Ross, L., M. Lepper, and M. Hubbard. 1975. Perseverance in self-perception and social
perception: Biased attribution processes in the debriefing paradigm. Journal of Personality and Social Psychology 32 (5): 880-892.
Rothschild, M. and Stiglitz, J. 1976. Equilibrium in competitive insurance markets: An essay on
the economics of imperfect information. Quarterly Journal of Economics 90 (4): 629-649. Sarbanes-Oxley Act. 2002. Public Law No. 107-204. Washington, DC: Government Printing
Office. Savage, L. 1954. The Foundations of Statistics. New York: Wiley. Schleifer, A. 2000. Inefficient Markets: An Introduction to Behavioral Finance. Oxford:
Oxford University Press. Schmidt, F., J. Hunter, and A. Outerbridge. 1986. Impact of job experience and ability on job
knowledge, work sample performance, and supervisory ratings of job performance. Journal of Applied Psychology 71 (3): 432-439.
Schrand, C. and B. Walther. 2000. Strategic benchmarks in earnings announcements: The
selective disclosure of prior-period earnings components. The Accounting Review 75 (2): 151-177.
Seifert, C., A. Patalano, K. Hammond, and T. Converse. 1997. Experience and Expertise: The
role of memory in planning for opportunities. In Expertise in context, eds. R. Feltovich, K. Ford, and R. Hoffman, 101-123. Menlo Park, CA: The MIT Press.
Serwer, A. 1998. It’s the end of the world as we know it, and I feel fine. Fortune (September
28): 317.
52
Shiffrin, R. and W. Schneider. 1977. Controlled and automatic human information processing.
II. Perceptual learning, automatic attending and a general theory. Psychological Review 84 (2): 127-190.
Silverman, R. 2002. GE to change its practices of disclosure. The Wall Street Journal, February
20: A3. Simon, H. 1979. Models of Thought, Vol. 1. New Haven, CT: Yale University Press. Sinha, P., L. Brown, and S. Das. 1997. A re-examination of financial analysts’ differential
earnings forecast accuracy. Contemporary Accounting Research 14 (1): 1-42. Snyder, M. and W. Swann. 1978. Behavioral confirmation in social interaction: From social
perception to social reality. Journal of Personality and Social Psychology 36 (11): 1202-1212.
Spence, M., 1973. Job market signaling. Quarterly Journal of Economics 87 (3): 355-374. Sternberg, R. 1997. Cognitive conceptions of expertise. In Expertise in Context, eds. P.
Feltovich, K. Ford, and R. Hoffman, 149-162. Menlo Park, CA: The MIT Press. Sternberg, R. and E. Grigorenka. 2003. The psychology of abilities, competencies, and expertise.
New York: Cambridge University Press. Stewart, T., P. Roebber, and L. Bosart. 1997. The importance of the task in analyzing expert
judgment. Organizational Behavior and Human Decision Processes 69: 205-219. Stickel, S. 1992. Reputation and performance among security analysts. Journal of Finance 47
(5): 1811-1836. Tesser, A. 1978. Self-generated attitude change. In Advances in Experimental Social
Psychology: Volume 11, ed. L. Berkowitz, 288-338. New York: Academic Press. Thaler, R. 1999. The end of behavioral finance. Financial Analysts Journal 55: 12-17. Tuttle, B., M. Coller, and F. Burton. 1997. An examination of market efficiency: Information
order effects in a laboratory market. Accounting Organizations and Society 22 (1): 89-103. Tversky, A. 1969. Intransitivity of preferences. Psychological Review 76: 31-48. Tversky, A. and D. Kahneman. 1981. The framing of decisions and the rationality of choice.
Science 211 (4481): 453-458. Tversky, A. and D. Kahneman. 1986. Rational choice and the framing of decisions. Journal of
Business 59 (4): 251-278.
53
Tversky, A. and D. Kahneman. 1991. Loss aversion in riskless choice: A reference-dependent model. Quarterly Journal of Economics 106 (4): 1039-1061.
Unnava, H., and R. Burnkrant. 1991. Effects of repeating varied add executions on brand name
memory. Journal of Marketing Research 28 (4): 406-416. Veverka, M. 1999. While Amazon.com spends like crazy, its customers begin to trim their
expenditures. Barrons (August 2): 47. Von Neumann, J., and O. Morgenstern. 1947. Theory of Games and Economic Behavior.
Princeton, NJ: Princeton University Press. Watts, R., and J. Zimmerman. 1997. Positive Accounting Theory. Englewood Cliffs, NJ:
Prentice-Hall. Williams, P. 1996. The relation between a prior earnings forecast by management and analyst
response to a current management forecast. The Accounting Review 71 (1): 103-115.
54
FIGURE 1 Theoretical Basis and Methodology
of Recent Empirical Publications on Financial Accounting Topics
Economic Theory
Psychology Theory
Total
Archival Method 580 articles (98%)
13 articles (2%)
593 articles
Experimental Method 10 articles (29%)
25 articles (71%)
35 articles
Total 590 articles
38 articles
_________________________ This figure is based on empirical (i.e., data-based) financial accounting papers published during the 1993-2003 time period in the following journals: The Accounting Review, Contemporary Accounting Research, Journal of Accounting Research, Journal of Accounting and Economics, and Review of Accounting Studies. We categorized each study based on its methodology and theoretical basis. Specifically, the studies were classified based on whether they relied on archival data or data created for the study using an experiment or survey. Each study also was classified based on whether the theory underlying the study’s predictions was primarily psychology-based or economics-based.
55
FIGURE 2
Prospect Theory’s Value and Probability Weighting Functions
Panel A: Value Function
Panel B: Probability Weighting Function
Source: Kahneman and Tversky (1979)