Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Incorporating social preferences into incentive system design
(USC Thesis Other)
Incorporating social preferences into incentive system design
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
INCORPORATING SOCIAL PREFERENCES INTO INCENTIVE SYSTEM DESIGN
by
Melissa Martin
____________________________________________________________________
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
BUSINESS ADMINISTRATION
December 2008
Copyright 2008 Melissa Martin
ii
Acknowledgements
I thank my dissertation committee Kenneth A. Merchant (chair), Wim A. Van der
Stede, and Peter Kim for their endless support. I thank the Institute of Management
Accountants for their generous support on this project. I also thank Tatiana Sandino,
S. Mark Young, and Sarah Bonner for the help they have provided throughout my
medical practitioneral program. Finally, I thank the workshop participants at the
University of Southern California, Arizona State University, Harvard Business
School, University of Notre Dame, George Mason University, Northeastern
University, Naval Post-graduate School, and North Carolina State University for
their helpful comments.
Additionally, I would like to thank the executives at my research site for their hours
of endless guidance, education, and support. The employees at this site also
generously gave their support and time for data collection and education.
Finally, I would like to thank my family for their support and guidance throughout
my graduate studies. Without the support of my husband, both emotional and
financial, I am not sure I would be here today.
Thank you to everyone.
iii
Table of Contents
Acknowledgements ii
List of Tables iv
Abstract v
Chapter 1. Introduction 1
Chapter 2. Literature Review 6
2.1 Importance of Incentives 6
2.2 Economic Theory of Incentives 6
2.3 Effectiveness of Incentives in Empirical Tests 8
2.4 Testing Economic Assumptions 12
2.5 Social Preferences 16
2.6 Inequity Aversion 21
Chapter 3. Evidence from the Field: The Case of ABC Health 23
3.1 Hypotheses 23
3.2 Research Site 29
3.3 Sample, Variables, Research Design 31
3.4 Survey Analysis 34
3.5 Results 35
Chapter 4. Conclusion 49
4.1 Discussion 49
4.2 Firm Outcomes 50
4.3 Limitations 51
4.4 Contributions 52
Bibliography 54
iv
List of Tables
Table 1: Survey Statistics 36
Table 2: Descriptive Statistics 37
Table3: Univariate Correlations 38
Table 4: Regressions of Production on Pay Type 40
Table 5: Regressions of Production on Incentive System Fit Proxy 41
Table 6: Regressions of Production on Incentive System Fit by High/Low
Performers 42
Table 7: Regressions of Production on Inequity Aversion 43
Table 8: Regressions of Production on Inequity Aversion with Additional
Controls 44
Table 9: Regressions of Production on the Inequity Aversion and
Subjectivity/Objectivity Interaction 46
Table 10: ANCOVA Test of the Inequity/Subjectivity Interaction 47
v
Abstract
This study examines whether the inclusion of social preferences in the assumed
utility function of individuals can provide a richer understanding of the incentive-
performance relationship. Recent economic research has shown that individuals’
utility functions consist of not only their own wealth and leisure, but also social
preferences such as reciprocity, honesty, and inequity aversion. This paper tests
whether incentive systems, all employed within a single firm, have differing effects
on performance when their fit with several aspects of utility is considered. The
results show that an increased fit between incentive systems and individual
preferences is associated with increased performance. Additionally, when employees
exhibit high levels of a specific social preference -- aversion to differences in pay --
performance suffers under a pay-for-performance scheme. Finally, when an
employee exhibits high levels of aversion to differences in pay and is under a
subjective incentive scheme, the negative performance effects are exacerbated. This
research contributes to the field by broadening our understanding of incentive
systems by further understanding the factors which impact the effectiveness of the
incentive-performance relationship.
1
Chapter 1: Introduction
Management accounting researchers, as well as managers, have long
recognized the value in aligning the incentives of employees with the goals of the
firm. The information provided by complex management accounting systems has
allowed for a substantial increase in the use of monetary incentives as perhaps the
single most important tool to achieve this incentive alignment and, thereby, to
increase the performance of employees (Cole, 1986). As employee compensation
ranks as one of the largest costs of doing business, at 56.4% of the total value of
goods and services produced in the U.S. in 2006 (not including employee benefits)
1
,
the question of determining how to create the most “bang for your buck” is of great
importance to both academics and practitioners (U.S. Bureau of Economic Analysis,
2006). The need to align employee behavior and performance to business goals is the
key driver “prompting 87% of companies to use incentive compensation programs”
(Jutras, 2007, p. 1). But, “strangely, in spite of the large amounts of money spent and
the obvious relevance of behavioral theory for industrial compensation practices,
there is probably less solid research in this area than in any other field related to
worker performance” (Opsahl and Dunnette, 1966, p. 94).
Economic contracting theory contends that performance can be maximized
when the goals of the employee are more closely matched with those of the firm.
This idea was popularized through agency theory, which utilizes the simplifying
assumption that people are purely self-interested; i.e., they have utility only for their
1
Non-salary compensation composes 19% of this total compensation cost.
2
own wealth and leisure (Baiman, 1982, 1990). These theories assert that employees
must be compensated to accept the disutility associated with increased effort; this
ensures increased motivation and thereby output. However, empirical evidence on
performance-based incentives does not always substantiate these claims. There is
evidence both supporting the effectiveness of incentives (e.g., Banker et al., 1996a;
Lazear, 2000) and questioning it (e.g., Lee et al., 1997; Wright and Kacmar, 1995) in
the accounting, psychology, and economics literatures (see Bonner, 2000; Jenkins,
1986; Jenkins et al., 1998 for reviews). In fact, a 2004 survey of 350 firms shows
that 83% of organizations believe that their pay-for-performance systems are “not at
all” or only “somewhat” effective (Pfeffer and Sutton, 2006). This disparity in the
theoretical predictions and empirical evidence brings into question the effectiveness
of financial incentives.
Economists and psychologists alike (e.g. Rabin, 1993; Fehr and Schmidt,
1999; Andreoni and Miller, 2000; Falk et al., 2000; Evans et al., 2001; Luft, 1997)
have begun to realize that the utility function including only an individual’s own
wealth and leisure is not only a simplified assumption, but one that can lead to
incorrect behavioral predictions. Theoretical economists also argue that changes in
the utility function can “have a decisive impact on how material incentives are
constituted and how they function” (Fehr and Fischbacher, 2002, p. C1-C2).
Furthermore, Fehr and Falk contend that, “progress in understanding the actual
effects of incentives has been limited by constraining attention to an empirically
questionable view of human motivation”, as in the “self-interest” hypothesis (2002,
3
p. 719). While the intuition behind agency and incentive theories is not questioned,
their use of this simplifying assumption may lead to incorrect behavioral predictions
under some circumstances. As such, economists have begun to reconsider this
assumption and incorporate social preferences into the simplified utility function
(e.g., Guth et al., 1982; Roth et al., 1981). Social preferences are defined as
preferences not only for the material resources (compensation) allocated to the agent,
but also for the material resources allocated to relevant reference agents (Fehr and
Falk, 2002).
Taken together, the experimental and theoretical economics literatures
maintain that people have preferences for more than wealth and leisure alone and
that these other-regarding preferences can differ across people and across time. More
importantly, theoretical economists also argue that these alterations in the economic
theory, particularly changes to the assumed utility function, can have a decisive
impact on how incentive systems should be structured and how they function (Fehr
and Fischbacher, 2002). This change to the underlying economic assumptions may
provide a richer understanding of incentive systems. More specifically, generic
incentive programs, those that do not take into account the differences in
preferences, both individualistic and social, can fail if the effects of these
assumptions are not considered (Fehr and Fischbacher, 2002).
This study argues that the greater the fit between the incentive system and the
employees’ preferences, the higher the level of individual performance. The results
support this prediction of a direct relationship between this specific measure of
4
incentive system fit and resulting individual performance. The findings also show
that employees with higher levels of aversion to compensation differences
2
, a
specific social preference, perform significantly worse than their individualistic
counterparts under an incentive compensation scheme, even after their own
performance is controlled. Finally, the results show that this aversion to
compensation differences interacts with specific aspects of the compensation system
(e.g., amount of subjectivity) to produce differing performance results. Specifically,
the study finds that subjectivity exacerbates the negative performance effects of
inequity aversion.
This study adds to the incentives literature by demonstrating empirically the
importance of considering social preferences in incentive system design. Recent
research in accounting has argued for the existence of social preferences (Hannan,
2000; Luft, 1997). However, the existing empirical evidence is limited in scope and
the interactions of social preferences and the heterogeneity of participants have not
been considered. This study uses data from a unique real-world setting capable of
testing these social preference effects. This study differs from prior research in that
it, first, uses a unique dataset to test the effects of different incentive systems within
the same firm. One feature of this data is that employees of the same level,
performing the same job, within the same firm are paid using differing incentive
systems. This allows for significant control of important aspects of the research,
particularly the holding constant of person, task, and environmental variables
2
This specific social preference was chosen due to its prevalence within the firm’s normal operations.
See Section 3.2 for more detail.
5
including ability, skill, and the corporate structure across employees. Second, the
accounting literature has substantially focused on CEO/executive incentives and firm
performance to test the relationship between incentives and performance, with a few
key exceptions (e.g., Anderson et al., 2007; Banker et al., 1996a; Banker et al.,
1996b, 2000; Core and Guay, 2001; Matsumura and Shin, 2006). This study
addresses one fundamental issue of incentive research in that it examines individual
incentives and the resulting individual performance. Also, this study tests the effects
of incentives at a lower level employee rank, as called for by recent accounting
research (Bol, 2007; Franco et al., 2007; Indjejikian, 1999).
6
Chapter 2: Literature Review
2.1: Importance of Incentives
Employee compensation is one of the largest costs of doing business in the
United States. In 2006, the cost of employee compensation (not including employee
benefits) totaled 56.4% of the total value of goods and services produced during the
same year (U.S. Bureau of Economic Analysis, 2006). A survey done by Baiman
(1982) shows that in 1982 only 30% of firms compensated employees using financial
incentives. Ledford et al. (1995), in a survey of Fortune 1000 firms, determined that
between 1987 and 1993 the instance of firms using of incentive-based pay for at least
20% of its employees increased from 38-50%. Denis et al. (2006) support the claim
of widespread incentive use by showing that in 2001 incentive compensation
comprised nearly two-thirds of the total compensation cost of the average firm. By
2007, Jutras (2007) claims that over 87% of firms use incentive-based compensation,
with over 67% of professionals within these organizations receiving bonus and/or
incentive-based compensation.
2.2: Economic Theory of Incentives
Some of the early academic research on incentives was theoretical in nature
and attempted to argue how and why monetary incentives would impact
performance. One proposition underlying the fundamental theoretical basis for
incentives is principal-agent theory (Baiman, 1982, 1990; Jensen and Meckling,
1976). First described in the accounting literature by Demski and Feltham (1978),
this theory attempts to describe the relationship between a principal who assigns
7
work to an agent “using the metaphor of a contract” (Eisenhardt, 1989). Using the
contract as the unit of analysis, this theory attempts to establish the most efficient
contract directing the principal-agent relationship given several assumptions (e.g.
self-interest, bounded rationality, and risk aversion) (Eisenhardt, 1989). This theory
has led to empirical research testing several aspects of the principal-agent theory,
including the use of performance-contingent compensation.
In its most basic form, this theory uses the broad economic simplifying
assumption that all individuals share preferences based solely on their own wealth
and leisure. That is, all individuals are purely self-interested. Formally, let (1,2,…n)
denote a set of individuals. In addition, x = (x
1
,x
2
,…x
n
) denotes the allocation of
physical resources where x
i
is the material resources (compensation) allocated to
person i and e
i
is the disutility for effort. The self-interest hypothesis then states that
for all individuals:
U
i
=
i
(x
i
, e
i
)
Where, U
i
is the utility level of individual i. This equation quantifies the theory that
individuals gain utility from their own monetary resources, and disutility from the
effort they must exert to obtain them. The inference, therefore, is that the agent will
shirk unless provided sufficient incentives (Minkler, 2004). These standard principal-
agent models argue that because an agent’s effort often cannot be readily monitored,
contracts should be written on all aspects of the agent’s performance that are
verifiable and informative of that effort (Jensen and Meckling, 1976). These
8
contracts contain material incentives which motivate employees to produce high
effort in exchange for increased wealth.
2.3 Effectiveness of Incentives in Empirical Tests
This theoretical literature asserts that the institution of properly written
material incentive contracts (those where effort can be observed) will consistently
lead to increased performance on the part of the agent. However, the empirical
literature in several fields has found evidence both supporting and questioning these
claims. These studies not only span decades, but also the literatures of psychology,
economics, management, accounting, and finance. Within accounting and economics
several studies have shown support for the effectiveness of performance-based
incentives (Banker et al., 1996a; Lazear, 2000; Tuttle and Burton, 1999; Fernie and
Metcalf, 1996; Paarsch and Shearer, 1999; Groves et al., 1994).
In one of the first studies to examine front-line worker’s incentive contracts,
Banker (1996a) finds that the introduction of incentive compensation for sales
employees induces a sales increase that is persistent and increases over time. In
Tuttle & Burton (1999), subjects suggest stock prices using several different cues for
firm performance. These results suggest that monetary incentives lead to increased
response times and higher levels of information usage, their measures of increased
performance. Lazear (2000) explores what happens to the performance of auto-glass
workers when the compensation system is converted from hourly wages to piece-rate
pay. The author finds an increase in performance of 44%, of which half is attributed
9
to increased productivity per worker and half to self-selection of more productive
workers into the firm and less productive employees out of the firm.
Paarsch & Shearer (1999) study the use of piece-rate incentives for tree
planters in British Columbia. The authors find that the use of piece-rates relative to
salaries results in an upper bound to incentive improvement in productivity of 35%.
An additional study on the relationship between performance and incentive
compensation by Fernie and Metcalf (1996) shows that British jockeys who are only
paid when their races are won perform better than jockeys who are on fixed pay;
again supporting the effectiveness of incentives at generating increased performance.
Groves et al. (1994) explore the use of incentives in Chinese firms moving toward a
more capitalistic structure. The authors find that as the responsibility for output
decisions is shifted from the state to the firm and there is an increase in bonus
payments and contract labor, the productivity of the firm also increases.
In his 1990 study, Ashton (1990) finds that in the absence of decision aids
incentives lead to increased performance as measured by increased decision
accuracy. Awasthi and Pratt (1990), in a task involving students solving complex
accounting problems, conclude that incentives increase performance when perceptual
differentiation
3
is high. In a study involving a recall task, Kahneman and Peavler
(1969) ask their subjects to learn a series of number noun pairings. The subjects are
then asked to recall these pairings in differing incentive cells. The authors find that
subjects provide higher levels of monetary incentive outperformed their low-
3
Perceptual differentiation is defined as “an individual’s ability to perceptually abstract from a
complex setting certain familiar concepts or relationships”.
10
incentive counterparts 55% to 18% correct in recalling these pairings. Sprinkle
(2000) finds that incentives only generate positive performance effects after
significant experience and feedback. Finally, Libby and Lipe (1992) examine the
effect of incentives in the task of recognizing and recalling a list of internal
accounting controls. The authors conclude that subjects who received incentives,
relative to a flat wage, increased the duration spent on the task and this led to
improvements in both the recall and recognition task.
However, other studies question this direct incentive-performance
relationship by finding either no result from the implementation of financial
incentives or a negative performance effect (Ashton, 1990; Awasthi and Pratt, 1990;
Bouwens and van Lent, 2006; Fessler, 2003; Libby and Lipe, 1992; Sprinkle, 2000;
Hogarth et al., 1991; Lee et al., 1997; Weiner, 1966; Wright, 1989; Wright and
Kacmar, 1995). In Ashton (1990), the author establishes that when a decision aid is
present in the task of predicting bond ratings, incentives actually decrease decision
accuracy. In another accounting related task, Awasthi and Pratt (1990) find that
when perceptual differentiation is low, incentives have no effect on task
performance. In a field study of Dutch firms by Bouwens and van Lent (2006), the
authors suggest that any performance effects induced by incentives are mostly driven
by selection effects and are not related to increased individual performance. Fessler
(2003) determines that when a task was considered attractive prior to the start of the
experiment, financial incentives worsened performance. However, when the task
was considered unattractive there was no effect of financial incentives on
11
performance. In Sprinkle (2000) the author highlights the importance of his results
existing only in the multi-period setting. Hogarth et al. (1991) study the effect of
incentives in an environment which also includes punishment. The authors
determine that in an intermediate punishment environment financial incentives have
no effect on performance. However, in exacting punishment environments financial
incentives decrease subject’s performance.
Several other studies have found no relationship between financial incentives
and increased performance. For example, Lee (1997) has subjects solve four-digit
arithmetic problems. The results show no difference in performance between the
flat-rate and piece-rate groups. Weiner (1966) utilizes a recall task to test the
incentive-performance relationship and finds no difference between the two groups.
Finally, in a card-sorting task, Wright (1989), finds no difference in performance
between individuals who were paid a flat wage and those paid on a piece-rate
scheme.
In an attempt to consolidate the literature on financial incentives, which spans
many disciplines and decades, several authors have performed meta-analyses on the
results of the supporting studies. The first of these, conducted by Jenkins (1986)
examines 28 laboratory and field experiments and finds an overall positive
association between incentives and performance.
4
In 1998, Jenkins et al., undertook
a second review of the literature including studies published in the interim decade.
4
While this study attempts to determine the overall association between financial incentives and
performance, only one study, Toppen (1965), actually examines the direct relationship between
financial incentives and task performance.
12
This review, unlike the first which was substantially qualitative, is a quantitative
analysis not only of the incentive-performance relationship but also of some potential
moderators including, task type, setting, and theoretical framework. The results
show a correlation of .34 of incentives with performance quantity, but no
relationship with associated quality, and confirm setting and theoretical framework
(but not task type) as moderators.
The next major review of this literature was done by Bonner et al. (2000).
This study includes 131 laboratory studies which relate financial incentives and
resultant performance.
5
Overall, the study finds that in only about half of the cases
do financial incentives lead to increased performance. The study also examines two
potential moderators of the relationship: task type and incentive scheme type. The
results show that task type matters; as task complexity increases the relationship
between incentives and performance decreases. They also show that quota schemes
are the best at eliciting performance followed by piece-rate, tournament, and flat
incentive schemes. The above meta-analyses have attempted to broadly examine the
incentive-performance relationship; however, they have been restricted to laboratory
experiments.
2.4 Testing Economic Assumptions
The theory of incentives is based on economic assumptions presumed to hold
true in the general compensation environment. One of the main assumptions in
incentive theory is the self-interest hypothesis which claims that individuals are
5
This review does not include any field-based studies. See Prendergast (1999) for an explanation of
field studies exploring the incentive-performance relationship.
13
concerned only for their own wealth and leisure. In the early 1990’s experimental
economists began to doubt the validity of the self-interest hypothesis and began to
design experiments to validate this assumption (Berg et al., 1995; Cameron, 1999;
Fehr and Schmidt, 1999; Fehr and Tougareva, 1995; Guth et al., 1982). Through the
use of several types of experimental economics games, researchers in this field have
found many instances in which the assumptions of traditional economics do not hold.
One such game, invented by Guth (1982), is based on the idea that a pair of
subjects must agree on how to split a sum of money. In this Ultimatum Game the
first player (the proposer) makes an offer of how to split the sum to the responder.
The responder then has the option to accept or reject the proposal. If it is accepted
the proposal is put into play and the sum is divided as suggested; if it is rejected both
players receive nothing. Economic theory, and its assumptions, tells us that the
responder should be willing to accept any positive amount as this will make him
better off than when he started. The proposer should realize this and therefore offer
the smallest amount possible to the responder. However, the mean result shows that
when the offer made by the proposer is less than twenty percent of the total
endowment, the responder rejects the offer with probability ranging from 40-60%.
This result is inconsistent with the self-interest hypothesis which suggests little
transfer and no rejection.
Another game used extensively throughout the literature is the Public Goods
Game (see Ledyard (1995) and Dawes and Thaler (1988) for reviews). The premise
for this game is that some number of players must decide how much of their
14
endowment to donate to a public good. This setup ultimately results in the player
receiving his own endowment less his contribution to the public good plus some
return based on the contributions of all other players to that good. Given the
assumptions of economics, the dominant strategy in this game is to contribute
nothing to the public good. The results regularly show that some people routinely
give some amount greater than zero in the first periods of the games, but in the last
period at least 75% of the players give nothing to the good. This result suggests that
the players are not behaving in a purely self-interested manner, especially in the
early periods.
A common addition to the traditional Public Good Game is the inclusion of
the opportunity to punish other players in the game (Ostrom et al., 1992). In this
form of the game the same format is kept for the initial stage. The change comes
when the participants are offered the chance to know how everyone contributed to
the good in part one. The players can then assign up to 10 punishment points to each
other player in the game. Assigning these points however is costly, both to the
punisher and to the punished player. The punisher is charged against their part one
income based on a monotonically increasing convex cost schedule, while the person
being punished is charge three points on average. Since this punishment is costly to
the punisher, the self-interest hypothesis would suggest that no player would ever
punish another. The rationality of the other players would then suggest, as in the
traditional game, that no contribution to the public good is the optimal play.
However, the introduction of the punishment feature results in subjects cooperating
15
at the astounding rate of 75% of their initial endowment. This result shows that
subjects feared the punishment of the other players, even though punishment is at
odds with the traditional economic assumptions.
In the Gift Exchange Game, first done by Fehr et al. (1993), the proposer
offers some amount to the responder which can be viewed as a wage payment. The
responder then has the opportunity to accept or reject this wage (as in the Ultimatum
Game). If the responder rejects the wage both players receive nothing; however, if
he accepts the wage he must then decide on some level of effort to return to the
proposer. In this game the payoff to the proposer is then the marginal value of this
effort less his chosen wage payment, while the responder receives the wage payment
less the cost of that effort (given in a scale during the experiment). Under the
standard economic assumptions, the responder will always give the lowest level of
effort possible and will never reject an offer (since this will always make him better
off than he was). Knowing this, the proposer will always offer the lowest possible
wage.
Since this game accommodates many of the features of the model of agency
theory including the highly incomplete contracts it has been tailored to fit many
variations (Charness, 2004a; Hannan et al., 2002; Fehr and Tougareva, 2002;
Heijden et al., 2001; Brandts and Charness, 2004). The overall result in these games
shows a consistent correlation between the wage offer made and the effort returned.
This implies that the responders are rewarding the generous offer made by the
proposer with generous effort. However, these studies demonstrate the heterogeneity
16
of the players in that some show no discernible pattern in their wage-effort
relationships. Some always act in a purely selfish manner, while others offer some
form of reciprocal effort.
In addition to these three games, two additional games have been frequently
used in the experimental economics literature, the Dictator Game and the Trust
Game. The Dictator Game, first introduced by Forsythe et al. (1994), operates in the
same manner as the Ultimatum Game, however the responder cannot reject the offer
of the proposer. Results from this game show that the proposers make lower offers
on average than in the Ultimatum Game; implying that they expect some level of
rejection in the low levels of their offer. Berg, Dickhaut, and McCabe (1995), first
introduced the Trust Game in which the proposer gets some amount of money from
the experimenter. This subject can then choose any amount of this money to forward
to the responder. In doing this, the responder will receive a value of money equal to
three times what was sent to him by the proposer. The responder then has the choice
how much, if any, of the money to return to the proposer. The results from these
studies again show that a correlation exists between what the proposer initially sends
and what is returned by the responder. However, these results (along with results
from the Dictator Game) are at odds with the predictions based on the self-interest
hypothesis.
2.5 Social Preferences
The mixed empirical evidence regarding incentives paired with experimental
economics evidence contrary to the self-interest hypothesis (e.g., Guth et al., 1982;
17
Roth et al., 1981)
6
has encouraged theoretical economists to reconsider the use of the
simplifying “self-interest” assumption in the standard utility function. “This
assumption is parsimonious and facilitates analysis; however, while it is generally
useful and often innocuous, it leads to unrealistic predictions of behavior in some
economic situations” (Charness, 2004b), p.665). These economists argue that adding
social preferences to the utility function can increase its generalized predictive
power.
An individual exhibits social preferences if he cares not only about the
material resources allocated to him, but also about the material resources allocated to
relevant reference agents (Fehr and Falk, 2002). Formally using the same notation as
above, the utility function which incorporates these social preferences is defined as
follows:
U
i
=
i
(x
i
, x
j
,
e
i
), where i ≠ j
Hence, a person exhibits social preferences if his/her utility function includes not
only his/her own monetary payoff x
i
, but also the monetary payoffs of others, x
j
.
Social preferences can include preferences for reciprocity (e.g., Rabin, 1993),
inequity aversion (e.g., Fehr and Schmidt, 1999), pure altruism (e.g., Andreoni and
Miller, 2000), spiteful or envious behavior (e.g., Falk et al., 2000), honesty (e.g.,
Evans et al., 2001), and fairness (e.g., Luft, 1997). Reciprocity has been studied
frequently within a laboratory setting. For example, Charness and Rabin (2002)
6
These studies use laboratory games such as the Ultimatum Game and Public Goods Game to infer
the existence of social preferences. The studies also indicate that the social preferences are not an
artifact of the lab setting (small monetary pay-outs) by testing the games in other settings where the
pay-outs represent a substantial monetary payout to the participants.
18
examine the effects of reciprocity in experimental games conducted in both
Barcelona, Spain and Berkeley, California. The experimental procedure utilized
simple games in which two participants were given several monetary choices. For
example in one game, Player B was given the choice between two pay-off sets. The
first set is a pay-off of 400 to both Player A and B (400,400); the second set is a pay-
off of (750,400) or 750 to Player A and 400 to Player B. The other instances of the
game allow Player A to choose between a pay-off option (750,0) or allowing Player
B to choose between (400,400) and (750,400). In several iterations of the games, the
authors find that in games where Pareto-damage is allowed, when Player A has
helped Player B, or has had no choice, Player B responds with a low likelihood of
helping Player A. However, when Player A has hurt Player B there is a 19% chance
that Player B will sacrifice his own income to hurt Player A. These results show that
people behave in a manner inconsistent with the self-interest hypothesis. In these
games the participants are willing to sacrifice their own income to exhibit reciprocal
behavior.
The social-preference of altruism has again been studied using experimental
lab games. Andreoni and Miller (2000) explore the existence of altruism by
employing a modified dictator game. In this game, the subjects (advanced
economics students) are presented with a set number of tokens. The subjects must
then make the decision of how to allocate these tokens to themselves and another
participant. In different plays of the game the cost structure for holding/passing
tokens changes. For example, in one structure the subject can earn one point for
19
holding a token, but 3 points for passing it to the other participant. The results of the
study first indicate that subjects are behaving rationally given their social-
preferences 98% of the time; a finding which denies the claim that social-preference
behavior is irrational. The authors do find that some subjects, about 50%, behave in a
self-interested manner throughout the experiment. However, they also conclude that
altruism exists within the game; that is, at least some participants allocate tokens to
the other participant even when it is costly to them.
Spiteful or envious behavior has been suggested as an alternative explanation
for departure from the self-interest hypothesis in the ultimatum game. Kirchsteiger
(1994) uses results from previously conducted and published ultimatum games to
imply support for his theory of envious behavior. The ultimatum game allows a
proposer to make a suggestion on how to split a sum of money. If the responder
agrees the sum is split as proposed. However, if they do not, neither party receives
anything. This theory argues that the behavior observed in ultimatum games is driven
by the proposer’s fear of rejection of their offer by envious responders and not by
their preference for fairness.
Evidence for the social preference of fairness can be found in experimental
games as far back as the 1960’s. In his 1993 paper which formalizes the theory on
fairness, Rabin (1993) analyzes the results from other’s experimental ultimatum
games. These results show that quite frequently when two subjects participate in the
ultimatum game, their behavior indicates a preference for fairness and a departure
from the self-interest hypothesis. This can be seen first by the behavior of
20
responders who reject offers of some money (thus making them better off than they
were before) because they feel the offers were unfair. A second behavior which
indicates fairness can be seen on the part of the proposers who often offer the
responders more than the minimum required amount (thus coming out of their own
pocket) in order to appear fair.
The preference for inequity aversion, introduced by Fehr and Schmidt (1999)
stems from the idea that at least some agents do not like inequity in the distribution
of compensation among their reference group. Several studies in economics (Kagel
and Wolfe, 2001, Andreoni et al., 2003) have focused on testing the theory of
inequity aversion and have found a large number of subjects behavior to be
consistent with this theory.
7
A final social preference, which has been explored in the accounting
literature, examines the preference for honesty. Hannan et al. (2006) examine
managers’ preferences for honesty in an experimental setting. Here the authors use
business students to participate in an experiment which utilizes a participative budget
framework. In this setting the manager has some incentive to lie as this results in
increased managerial wealth. However, even with this incentive some managers
displayed behavior consistent with a preference for honesty even when it is costly to
them.
7
Please see section 2.6 for additional details.
21
2.6 Inequity Aversion
Inequity aversion is illustrated when at least some agents do not like inequity
in the distribution of compensation among their reference group (Fehr and
Fischbacher, 2002). Stated differently, “the theory of inequity aversion assumes that
some but not all agents suffer a utility loss if their own material payoffs differ from
the payoffs of other agents in their reference groups” (Bartling and von Siemens,
2004, p. 1). The Fehr & Schmidt (1999) model for inequity aversion formally defines
the inequity aversion model (simplified for two agents) as follows:
U
i
(x) = x
i
– α
i
max │x
j
– x
i
, 0│- β
i
max │x
i
– x
j
, 0│+ e
i
, where i ≠j, β
i
≤ α
i
,
and 0 ≤ β
i
< 1
This model contends that a person gets utility first from their own material resources
(compensation), x
i
. They then consider the resources allocated to referent others, x
j
.
Their own utility is decreased by any difference in the resource allocation between
themselves, x
i
, and these referent others, x
j
. However, utility is diminished more
when the others pay is larger than their own ( α
i
), than when it is less ( β
i
).
This model led to the development of several studies which have attempted to
either test the validity of the model (Andreoni et al., 2003; Kagel and Wolfe, 2001)
or to incorporate its theory into related questions (Bandiera et al., 2005; Englmaier
and Wambach, 2005; Grund and Sliwka, 2002; Rotemberg, 1994). Both Kagel &
Wolfe (2001), and Andreoni et al. (2003), conducted studies which found support for
the theory of inequity aversion with large numbers of subjects exhibiting behavior
consistent with its predictions.
22
Englmaier & Wambach (2005) test optimal incentive contracts under the
presence of inequity aversion. They find that when agents are inequity averse,
contracts must balance three main factors: insurance, incentives, and fairness. They
also suggest that when agents are inequity averse, principals may be able to use this
aversion as an incentive by paying more fairly as opposed to paying more. Finally,
the authors suggest that inequity aversion may be a basis for the widespread use of
team-based incentives.
Bandiera, Barankay, & Rasul (2005) study performance under the theory of
inequity aversion. Using personnel data from a large farm in the United Kingdom
the authors find that production is on average fifty percent higher under a piece-rate
scheme then under relative pay. This differential performance suggests that inequity
aversion plays a part in worker performance. In an employee survey piece,
Rotemberg (1994) finds that workers are not only concerned with the general profits
of the firm, but also with the distribution of these profits to others. Finally, Grund &
Sliwka (2002) show that inequity aversion among agents can lead to increased effort
when a tournament incentive scheme is utilized.
23
Chapter 3: Evidence from the Field: The Case of ABC Health
3.1 Hypotheses
Management accounting lists two main functions for the use of information
created by its systems: decision influencing and decision facilitating (Demski and
Feltham, 1976). The decision influencing role of management accounting strives to
provide information which can influence the decisions of managers to be more in
line with the goals of the owners (Vroom, 1964). This is often achieved by linking
performance to pay. A survey done by Baiman (1982), shows that in 1982 only 30%
of firms compensated employees using financial incentives. This number climbed to
over 90% of Fortune 1000 firms in a 1995 study by Ledford, et al. (1995).
Some of the early academic research on incentives was theoretical in nature
and attempted to argue how and why monetary incentives would impact
performance. One proposition underlying the fundamental theoretical basis for
incentives is principal-agent theory (Baiman, 1982, 1990). In its most basic form,
this theory uses the broad economic simplifying assumption that all individuals share
preferences based solely on their own wealth and leisure. These standard principal-
agent models argue that because an agent’s effort often cannot be readily monitored,
contracts should be written on all aspects of the agent’s performance that are
verifiable and informative of that effort (Jensen and Meckling, 1976). These
contracts contain material incentives which motivate employees to produce high
effort in exchange for increased wealth.
24
This theoretical literature asserts that the institution of properly written
material incentive contracts (those where effort can be observed) will consistently
lead to increased performance on the part of the agent. However, the empirical
literature in several fields has found evidence both supporting and questioning these
claims (e.g. Bonner et al., 2000; Berg et al., 1995; Cameron, 1999; Fehr and
Schmidt, 1999; Fehr and Tougareva, 1995; Guth et al., 1982). The mixed empirical
evidence regarding incentives paired with experimental economics evidence contrary
to the self-interest hypothesis (e.g., Guth et al., 1982; Roth et al., 1981) has
encouraged theoretical economists to reconsider the use of the simplifying “self-
interest” assumption in the standard utility function. These economists argue that
adding social preferences to the utility function can increase its generalized
predictive power.
An individual exhibits social preferences if he cares not only about the
material resources allocated to him, but also about the material resources allocated to
relevant reference agents (Fehr and Falk, 2002).
This theoretical model of social preferences led economists to question the
behavioral effects, if any, of the changes in the simplifying assumption of purely
self-interested preferences. The alterations in the assumptions of economic theory
can have a decisive impact on how incentive systems should be structured and how
they function (Fehr and Fischbacher, 2002). Some have argued that “material
incentives may reduce efficiency in situations in which they are predicted to be
efficiency enhancing by the self-interest model” (Fehr and Fischbacher, 2002, p.
25
C22). Theoretical economists also posit that incentives may have negative side
effects because they reduce a natural inclination for voluntary cooperation when
social preferences are present (Fehr and Fischbacher, 2002). Burks et al. (2006)
support this in their field study by showing that pay-for-performance schemes erode
cooperative behavior among bike messengers. Additionally, Fehr and Falk contend
that “economists may even fail to understand the effect of economic incentives on
behavior if they neglect these (social) motives. In particular, we will show that
because of the existence of these motives, economic incentives may backfire and
reduce the agent’s performance or compliance with rules” (2002, p. 688).
Rabin suggests that “the debate over whether there are systematic, non-
negligible departures from self-interest is over” (2002, p. 666). Results from
experimental economics suggest that these social preferences exist in a substantial
fraction of the population (Fehr and Fischbacher, 2002). “Moreover, since non-
pecuniary (social) motives interact in different ways with different types of
incentives, the neglect of these motives is also likely to create a distorted view of the
relative performance of these different incentives” (Fehr and Falk, 2002). The
theoretical work stresses the importance of aligning the incentive systems with the
preferences of the agents. An incentive system based on the premise of purely self-
interested agents will not have the desired effects in a population where at least some
agents have social preferences. It is therefore imperative for the firm to align the
incentive system with the heterogeneous preferences of its agents. This leads to the
following hypothesis:
26
H1: Incentive systems that better fit with the employee’s
preferences are associated with increased performance.
Inequity Aversion
“The theory of inequity aversion assumes that some but not all agents suffer
a utility loss if their own material payoffs differ from the payoffs of other agents in
their reference groups” (Bartling and von Siemens, 2004, p. 1). This inequity
aversion can then result in increased agency costs of providing incentives which can
sometimes “render flat wage contracts optimal even though incentive contracts are
optimal with selfish agents” (Bartling and von Siemens, 2004, p. 1). This model
suggests that agents can suffer some disutility from differences in compensation
between themselves and referent others. Agency theory then implies that this
disutility will be reflected in the behavior of the subordinate creating suboptimal
performance. This has been shown empirically by Pfeffer & Langton (1993) who
studied the performance, as measured by publications, of university faculty across
several departments. The authors concluded that as incentive programs differed
across individuals (as dispersion increased), the level of productivity decreased. This
finding provides empirical support for the theory presented above.
When incentives are introduced inequality in pay amount naturally arises
(Neilson and Stowe, forthcoming). At the research site not only do these inequalities
in pay amount naturally arise from the use of performance based incentives, but
inequalities are also created from differences in the structure of the incentive
27
programs. Nielson & Stowe more follow that “since there is a chance that workers do
care about how
8
and how much their coworkers are paid, it is important to derive the
correct incentive system in the presence of these preferences” (forthcoming, p. 1).
The use of multiple compensation schemes combined with the preference for
inequity aversion is then likely to affect the effectiveness of incentives. This leads to
the following hypothesis:
H2: In the presence of incentive compensation, individual
performance will decrease as the individual’s level of
inequity aversion increases.
It is likely, however, that an agent’s inequity aversion will affect his
performance differently depending on his perceptions of the compensation system.
Milgrom & Roberts add that “a given level of pay may be viewed as good or bad,
acceptable or unacceptable, depending on the compensation of others in the
reference group, and as such may result in different behavior” (1992, p. 119).
Another issue which may cloud the agent’s perception of the compensation system is
its use of subjectivity. Gibbs, et al. (2004) assert that subjectivity in compensation
systems plays an important role in these perceptions and the resulting behavior. They
also maintain that “we should not expect the use of subjectivity to always be
effective” (p. 410-411).
A subjective compensation scheme relies on some discretion and judgments
of a superior; while an objective system is put into play systematically across all
8
“How” inserted by author for clarity.
28
members. Thus, when differences in pay amount occur, they can be attributed to
different causes depending on the compensation system. In an objective system, such
as a defined piece rate contract, any differences in pay are attributed to differences in
the agent’s effort. Under a subjective system, differences in pay are attributed to
differences in the superiors’ judgments. For someone who strongly dislikes inequity
in pay, a subjective system may exacerbate performance problems since the agent
views the differences in pay as even more unfair because they are attributed to
subjective judgments. An objective system, in contrast, may mitigate the problems
by illustrating that differences in pay are earned and not merely received.
In their study of British farm workers, Bandiera et al (2005) find that under a
defined piece-rate scheme (objective) performance is fifty percent higher on average
than under a relative (subjective) scheme. This empirical result paired with the
theory above lead to the following hypothesis:
H3: The type of pay scheme interacts with an individual’s
preferences for inequity aversion to affect individual
performance. Specifically, when an individual’s inequity
aversion is high, individual performance under subjective
systems will be lower than under objective incentive
systems. When an individual’s inequity aversion is low,
there will be no difference in individual performance
under subjective and objective incentive systems.
29
3.2 Research Site
This research was conducted at a for-profit national health care firm
headquartered in Southern California. The firm, henceforth known as ABC Health,
has been a leader in its industry for nearly a decade. ABC Health is publicly traded
and owns approximately 450 operating units (clinics) across the US. The firm has a
market capitalization of about $3.3B and prior year revenues of almost $1B, placing
it securely within the S&P’s midcap group. The firm employs approximately 8,000
employees nationally, over 1,600 of which are medical practitioners (the sample
population). The advantage of using a research site such as this is that while access
has been granted to a single firm, the sample size is actually quite large considering
that each of the 450 units operates alone and each of the 1600 medical practitioner’s
production is tracked individually.
I entered the firm studying the relationships between several of their non-
financial performance measures and firm performance. However, the management
of the firm insisted that I spend a few weeks learning about the business and the
firm’s place within its industry. The operating strategy of ABC Health is to grow
through the acquisition of new operating units. One challenge with this strategy,
from a management standpoint, is that many of the systems in place at the clinic are
inherited upon purchase. The firm generally adopts the compensation scheme present
at the time of purchase,
9
while changing the employee structure to be consistent
9
The firm occasionally makes changes to the compensation systems in place at acquired clinics.
However, a significant portion of the compensation systems are left untouched.
30
across units. These differences in pay among employees of the same level struck me
as an incentive problem. This issue led me to develop the idea of inequity aversion.
Each unit has a designated medical director along with several medical
practitioners who serve under him/her. The pay structure for these medical
practitioners can vary to be any of the following: (1) straight salary, (2) salary or
commission
10
(whichever is greater), (3) straight commission, or a combination of
the above. A final layer of compensation differences comes by adding a bonus and/or
stipend program to any (or none) of the compensation schemes listed above. Having
these different pay schemes within the same firm, and level within the firm, creates a
unique research opportunity to test the differences in outcomes of these pay systems,
while other variables are held constant within the firm. This also provides the
advantage of being able to test the effectiveness of incentives at the individual level,
something rarely done in the accounting literature. In a similar vein, the data
provided by the firm is on lower-level employees, not the traditional CEO sample for
incentive research. Together, these qualities make ABC Health a unique research
setting that can contribute greatly to our understanding of performance-based
incentives and their behavioral effects.
10
Commissions also differ among medical practitioners. The percentage of the commission changes,
as well as the inclusion or exclusion of negative accruals. Negative accruals are administered by first
determining that for those medical practitioners who do not produce above the salary level, the salary
will be given for that pay period. However, if the inclusion of some salary based pay and some
commission based pay causes the individual to earn above the set commission based rate for the year,
deductions will be made to ensure the final pay will be the actual revenue production multiplied by
the set commission rate.
31
3.3 Sample, Variables, and Research Design
I collected data for this study both directly from the firm and through a
survey of its employees. Specifically, I obtained objective data regarding the
performance of the employees and units, as well as other non-financial variables
directly from the firm. The survey data consists of compensation, demographic, and
perceptual variables.
The data analysis and hypothesis testing use one major outcome variable. I
collected this performance-based outcome variable directly from the firm. The
variable, average monthly production (AVMOPROD), is the average revenue from
medical visits and procedures generated by each individual medical professional per
month. It is a revenue-based variable that reflects procedures controllable by the
medical professional. This is the only number used by the firm for compensation; no
cost or profitability measures are used.
11
The study uses two key independent variables: incentive system fit and
inequity aversion. To proxy for incentive system fit I use a variable from the
psychology literature -- satisfaction with pay (Lawler, 1971, 1973). Ideally, I could
have asked the medical practitioners direct questions regarding their perceptions of
the differences in the compensation systems across the firm. However, the
management of the firm was quite adamant that these direct questions would stir up
questions among the practitioners who would begin to talk about their compensation
schemes, something that was discouraged by management. This discussion with
11
As such, this study does not attempt to use cost or profit measures as indicators of performance.
32
management gave me anecdotal evidence that they believed that if the practitioners
were aware that other systems were available they may become even more
discouraged with their own.
The literature supports this proxy by arguing that the “perceived value of an
incentive lies in the extent to which it supports an individual’s goals” (Shah et al.,
1998, p. 285). That is, the degree to which an incentive is aligned with the
preferences of the individual can be seen in his attitudinal reaction to the incentive,
or his satisfaction with that pay. This construct has been developed by considering an
employee’s reaction to his/her own pay as well as the pay of referent others (social
preferences are considered). Heneman & Schwab (1985) constructed a twenty-item
instrument to measure the validity of these factors known as the Pay Satisfaction
Questionnaire (PSQ).
12
This questionnaire serves as the basis for the survey sent to
the employees of ABC Health. The satisfaction with pay variable (SATPAY) is
composed of eleven applicable questions. These questions demonstrate high
reliability with a Cronbach’s alpha of .92.
The second variable, inequity aversion, is defined as disutility stemming
from differences in the distribution of material resources (compensation) among
12
Heneman & Schwab proposed that the satisfaction with pay construct was a multidimensional
construct containing five factors: pay level, pay raises, benefits, structure, and administration. The
authors then constructed a twenty item instrument to measure the validity of these factors, the Pay
Satisfaction Questionnaire (PSQ). They administered the survey to over 350 individuals and used
factor analysis to determine the dimensionality of the results. Their results showed not five
dimensions, but four. The administration and structure dimensions were combined into a single factor
and two items were removed from the questionnaire. Since that time, several studies have attempted
to test the validity of the PSQ and of the idea that satisfaction with pay consists of the four dimensions
above. Judge (1994) reconciles the work in prior studies to support the initial results of Heneman and
Schwab which define four factors within the construct.
33
referent others (Fehr and Fischbacher, 2002). The inequity aversion variable is
composed of three questions from the employee survey (e.g. “How satisfied are you
with the consistency of the firm’s pay policies for medical practitioners?”). This
instrument shows inter-item reliability with Cronbach’s alpha of .64. These three
questions are averaged to form the inequity aversion variable (INEQAV).
I also measure two variables that are descriptive of the incentive system
design. The first variable (INCENTDUM) classifies employees as receiving either
fixed or variable (incentive) compensation. I created this variable using objective
data provided by the firm. The second situational variable (SUBJDUM) determines
which employees were contracted under a subjective versus objective compensation
scheme. I calculated this variable by multiplying the standard commission
percentage, twenty percent, by the medical practitioner’s actual annual production.
This number, the expected wage, was then compared to the actual wage received by
the medical practitioner. If the numbers differed by more than one standard deviation
in either direction the contract was classified as subjective.
13
Finally, I measured a number of control variables to help control for the
possibility of other explanations for differences in performance. Age (AGE) is a self-
reported variable collected directly from the employee survey. Location
(LOCATION) is a categorical variable denoting one of 6 regions to which the
medical practitioner belongs. This variable is included to control for the differences
13
It is clear that practitioners with different incentive rates will be considered subjective. This is the
classification desired; as I consider subjectivity to be the fact that one doctor was granted a
higher/lower rate than his peers.
34
in local demand and cost of living which may affect the outcome variable, annual
production. Gender (GENDER) is self-reported by each medical practitioner at the
time of the survey. Ethnicity (ETHNICITY) is a categorical variable self-reported by
each medical practitioner indicating their ethnicity. Years of experience as a medical
practitioner (YRSEXP) are also self-reported in the demographic section of the
survey. To control for differences in size among the operating units I use a variable
(NUMDOCS) which measures the number of full-time medical practitioners
employed at that operating unit. Additionally, since medical practitioner turnover has
occurred within the firm, the number of useable months of production differs. To
control for this, I use an objectively measured variable (MONTHS) which is the
number of months of production data provided for that individual medical
practitioner. Another concern is the effect of the actual wages on the employee’s
level of the two key variables listed above, incentive system fit and inequity
aversion. To control for this, I include the actual annual wages of each individual
medical practitioner (ANNWAGES). Similar to the month variable above, I also
included a control for the number of months of wage data which are provided
(WAGEMOS).
3.4 Survey Analysis
I sent the survey to 1,601 medical practitioners (the entire medical
practitioner population) employed by ABC Health as of June 1, 2007. These medical
practitioners range in experience from just one month to over 25 years. I mailed the
survey via US mail along with a cover letter and a postage-paid return envelope. The
35
cover letter detailed that the survey could be completed either on paper or using a
web-based survey form for which instructions were provided. The survey consists of
100 questions that address the satisfaction with pay variables, inequity aversion
variables, other general satisfaction questions, and demographic variables. Table 1
shows that of the 1601 surveys mailed out, 691 were returned by mail and 105 were
returned via the internet, generating an actual response rate of 50%. Of the surveys
returned, ten were blank and five had defaced the unique identifier, leaving a sample
of 779 useable surveys for a useable response rate of 49%. Responders were an
average age of 39.8 years, 36% male
14
, 74% were married, and the average
experience in the field was 13 years with an average of 4 years at ABC Health.
As a test for non-response bias, I use ANOVA to test the difference between
the early and late responders. I test the two key independent variables collected
through the survey, satisfaction with pay and inequity aversion, for differences
between the two groups. Table 1 (Panel C) shows that there is no significant
difference between early and late responders on either the satisfaction with pay
variable (F-statistic 0.09, p-value 0.77) or the inequity aversion variable (F-statistic
0.23, p-value 0.63).
15
Overall, there is no reason to believe that non-responders
would differ in any predictable manner from the responders.
14
This field is becoming substantially more female dominated. As of December 2005, there were
36,383 females in the field in the United States (compared with 43,186 men) making up almost 46
percent of the profession. The graduating class of 2007 (2,489 total students) is split 75.3 percent
female (1,873 students) and 24.7 male (616), according to the (McPheron, 2007).
15
These tests were also conducted on the mail and internet responder groups with no significant
differences found. Differences in demographic variables were also tested with no differences in
early/late or internet/mail responders found.
36
Table 1
Survey Statistics
Panel A shows the response rates, both actual and useable, for the medical practitioner survey. Additionally, non-responder
analysis is shown in Panels B and C. *** denotes significance at the 0.01 level (2-tailed). ** denotes significance at the 0.05
level (2-tailed). * denotes significance at the 0.10 level (2-tailed).
Panel A: Response Rate
Actual Useable
Total Surveys Sent 1601 1601
Received Paper 691 676
Received Online 105 103
Response Rate 0.50 0.49
Panel B: Dependent Variable Statistics
PAYSAT INEQAV
Mean Std.Dev Mean Std. Dev
Early Responders 3.34 0.85 1.90 0.73
Late Responders 3.32 0.85 1.94 0.73
Panel C: ANOVA Results
Early/Late
F-Stat Sig.
PAYSAT 0.09 0.77
INEQAV 0.23 0.63
3.5 Results
Table 2 shows that annual production averages $307,313 per medical
practitioner, while the median is slightly higher at $309,445. There is a mean of 7.3
medical practitioners at each clinic, and annual production is based on an average of
37
9.4 months
16
of production. Satisfaction with pay averages 3.03 out of 5.0 while
inequity aversion averages 2.97 with the median slightly higher in both cases. The
medical practitioners average wages of $45,379 annually, with the median annual
wage higher at $51,151 based on an average of 7.7 wage months.
Table 2
Descriptive Statistics
This table denotes descriptive statistics for dependent, independent, and control variable of interest in the study. I provide
variable descriptions below.
N Mean Std. Dev. Median
SATPAY 794 3.03 0.80 3.13
INEQAV 788 2.97 0.65 3.00
AVMOPROD 703 $32,159.28 $17,125.88 $32,408.50
ANNPROD 793 $307,313.37 $232,578.81 $309,445.00
NUMDOCS 793 7.30 7.94 4.00
MONTHS 794 9.35 4.35 12.00
ANNWAGES 793 $45,378.93 $39,217.10 $51,151.00
WAGEMOS 794 7.69 5.33 12.00
YRSEXP 762 12.68 10.42 10.00
GENDER 690 1.64 0.49 2.00
AGE 770 39.83 11.56 38.00
SATPAY = A survey response variable using a 5-point Likert scale. Measures the level of satisfaction with all aspects of
compensation, proxies for level of alignment between preferences and compensation scheme.
INEQAV = A survey response variable using a 5-point Likert scale. Measures the level of inequity aversion, distaste for
differences in compensation.
ANNPROD = The level of production (revenue generated from medical visits and procedures) generated by each individual
medical practitioner during the prior twelve months.
NUMDOCS = The number of medical practitioners employed at each facility. This variable serves to proxy for the size of each
operating unit.
MONTHS = The number of months production data was gathered for each medical practitioner during the prior 12 months.
ANNWAGES = The wages received by each individual medical practitioner over the prior twelve months.
WAGEMOS = The number of months of wage data gathered for each medical practitioner during the prior twelve months.
YRSEXP = The number of years of experience of each medical practitioner in the field.
AGE = The age of each medical practitioner self-reported at the time of the survey.
ETHNICITY = A categorical variable indicating self-reported ethnicity.
GENDER = A categorical variable indicating self-reported gender.
16
The firm recently underwent changes in the accounting system to calculate production for a number
of these operating units. Several months of production data were lost in the process. This, combined
with new hires, creates the average production months less than twelve.
38
Table 3
Univariate Correlations
This table denotes univariate Pearson correlations between dependent, independent, and control variables of interest. *** denotes significance at the 0.01 level (2-tailed). ** denotes significance at
the 0.05 level (2-tailed). * denotes significance at the 0.10 level (2-tailed).
INEQAV SATPAY ANNPROD MONTHS ANNWAGES WAGEMOS AGE GENDER AVMOPROD YRSEXP NUMDOCS
INEQAV 1.00
SATPAY -0.66*** 1.00
ANNPROD -0.05 0.00 1.00
MONTHS 0.07 -0.13*** 0.71*** 1.00
ANNWAGES 0.02 -0.06 0.76*** 0.57*** 1.00
WAGEMOS 0.03 -0.06* 0.76*** 0.70*** 0.85*** 1.00
AGE -0.02 0.03 0.15*** 0.19*** 0.07** 0.07** 1.00
GENDER 0.18*** -0.05 -0.10*** -0.05 -0.04 0.00 -0.42*** 1.00
AVMOPROD 0.21*** 0.18*** 0.90*** 0.14*** 0.57*** 0.46*** 0.06 0.11*** 1.00
YRSEXP -0.04 0.04 0.18*** 0.23*** 0.12*** 0.11*** 0.90*** -0.42*** -0.03 1.00
NUMDOCS -0.06 0.01 -0.14*** -0.15*** -0.09*** -0.12*** -0.08** 0.02 0.02 -0.07** 1.00
Note: Refer to Table 2 for additional variable definitions.
39
Table 3 reports Pearson correlations for control, dependent, and independent
variables of interest. The outcome variable, average monthly production, is
significantly correlated with several control variables, but not the two independent
variables indicating that multivariate tests to control for these relationships are
necessary.
Hypothesis Testing
I first analyze the effect of incentive compensation within the current firm
setting. I regress average monthly production on an indicator variable which
categorizes employees based on whether or not they receive any incentive-based
compensation (INCENTDUM) and controls (NUMDOCS, MONTHS, YRSEXP,
AGE, ANNWAGES, WAGEMOS, LOCATION, MARITAL, GENDER,
ETHNICITY). The results show that incentives are associated with increased
performance, measured as annual production. Medical practitioners receiving
incentive compensation outperformed their salaried counterparts by approximately
$4200 monthly.
Hypothesis one maintains that as the fit of the incentive system increases so
does the performance of the employee. To test this hypothesis, I regress average
monthly production on pay satisfaction (SATPAY) and controls (see above) for only
those practitioners who receive incentive compensation.
17
17
The regression sample contains only practitioners who receive incentive compensation for all
hypothesis testing.
40
Table 4
Regressions of Production on Pay Type
This table denotes the results of regressions of performance, as measured by average monthly production (AVMOPROD), on a
variable indicating fixed or incentive compensation (INCENTDUM). *** denotes significance at the 0.01 level (2-tailed). **
denotes significance at the 0.05 level (2-tailed). * denotes significance at the 0.10 level (2-tailed).
AVMOPROD
Coefficient (p-
value)
Intercept 32928.96
15.98***
INCENTDUM 4245.14
2.55**
NUMDOCS 303.25
4.85***
MONTHS -285.47
-1.71**
ANNWAGERES 12638.78
25.59***
LOCATION (GP2) -3796.61
-2.74**
GENDER 873.53
0.85
YRSEXP -29.03
-0.60
F-Statistic 69.58***
Adj. R
2
0.54
N 669
INCENTDUM = An indicator variable which equals one for all employees who receive incentive-based compensation.
LOCATION = A categorical variable indicating the one of 6 regions to which the medical professional belongs.
ANNWAGERES = This variable is the residuals from a regression of AVMODPROD on ANNWAGES. This variable is used
to control for endogeneity.
Note: Refer to Table 2 for additional variable definitions.
The results show a statistically significant relationship between annual production
and satisfaction with pay. This analysis indicates that a 10% increase in the pay
satisfaction results in increased monthly production of about $600 per medical
practitioner, or an aggregate annual increase of about $11.5 million. These results
41
Table 5
Regressions of Production on Incentive System Fit Proxy
This table reports the results of regressing performance, as measured by average monthly production (AVMOPROD), on a
proxy for incentive system fit, satisfaction with pay (SATPAY). *** denotes significance at the 0.01 level (2-tailed). **
denotes significance at the 0.05 level (2-tailed). * denotes significance at the 0.10 level (2-tailed).
AVMOPROD
Coefficient (p-value)
Intercept 29043.21
10.98***
SATPAY 1201.32
2.16**
NUMDOCS 331.99
9.39**
MONTHS -262.77
-1.55*
ANNWAGERES 12268.88
24.30***
LOCATION (GP2) -4323.96
-3.07**
GENDER 1314.93
1.27
YRSEXP -33.01
-0.67
F-Statistic 66.398***
Adj. R
2
0.53
N 618
Note: Refer to Table 2 for additional variable definitions.
support H1. It is possible, however, that individuals who are poor performers are also
the same individuals who are dissatisfied with their pay while good performers are
more satisfied with their pay overall. To control for the possibility of endogeneity
from actual performance, I run two additional regressions partitioned on actual
performance (Top/Bottom 20%). The results support H1 in both the high and low
performer groups.
42
Table 6
Regression of Production on Incentive System Fit by High/Low Performers
This table reports the results of regressions of performance, as measured by average monthly production (AVMOPROD), on
the incentive system fit proxy (SATPAY) partitioned by actual performance (Top/Bottom 20%). *** denotes significance at
the 0.01 level (2-tailed). ** denotes significance at the 0.05 level (2-tailed). * denotes significance at the 0.10 level (2-tailed).
AVMOPROD
Top 20%
AVMOPROD
Bottom 20%
Coefficient (p-value) Coefficient (p-value)
Intercept 41297.37 2912.83
5.58*** 1.26
SATPAY 383.82 1768.66
1.96** 3.88***
NUMDOCS 668.06 -67.04
5.41*** -1.26
MONTHS -307.01 286.08
-0.67 1.02
LOCATION (GP 2) -9332.52 -2506.99
-2.29** -2.32**
LOCATION (GP 3) 8888.69 4117.03
1.95** 1.98**
ANNWAGERES 6827.93 338.62
7.45*** 0.44
YRSEXP 18.60 45.53
0.14 1.02
F-Statistic 10.89*** 4.16***
Adj. R
2
0.49 0.24
N 112 154
Note: Refer to Table 2 for additional variable definitions.
H2 contends that performance will decrease as the level of inequity aversion
increases. I test this hypothesis by regressing average monthly production on the
inequity aversion variable (INEQAV) and the control variables. The results, a
43
Table 7
Regressions of Production on Inequity Aversion
This table reports the results of regressing performance, as measured by average monthly production (AVMOPROD) on the
aversion for inequity in pay (INEQAV). *** denotes significance at the 0.01 level (2-tailed). ** denotes significance at the
0.05 level (2-tailed). * denotes significance at the 0.10 level (2-tailed).
AVMOPROD
Coefficient (p-value)
Intercept 30129.37
10.02***
INEQAV -1108.70
-1.98**
NUMDOCS 334.67
5.40***
MONTHS -315.01
-1.85*
ANNWAGERES 12333.01
24.07***
LOCATION (GP2) -4090.73
-2.92***
GENDER 1553.73
1.43
YRSEXP -39.37
-0.79
F-Statistic 65.68***
Adj. R
2
0.52
N 617
Note: Refer to Table 2 for variable definitions.
significant negative coefficient on the inequity aversion variable, show that
production decreases as the level of inequity aversion increases, in support of H2.
The results are again economically significant. A 10% increase in inequity
aversion decreases monthly production by $554.35, or in aggregate slightly more
than $10.6 million annually. I run three additional regressions to control for other
44
Table 8
Regressions of Production on Inequity Aversion with Additional Controls
This tables reports the results of regressions of performance, as measured by average monthly production (AVMOPROD), on
the aversion to inequity in pay (INEQAV) while also controlling for other possible explanations of satisfaction with pay
(SATPAY) and the possible endogeneity of actual performance (Top/Bottom 20%). *** denotes significance at the 0.01 level
(2-tailed). ** denotes significance at the 0.05 level (2-tailed). * denotes significance at the 0.10 level (2-tailed).
AVMOPROD
AVMOPROD
Top
20%
AVMOPROD
Bottom 20%
Coefficient (p-value)
Coefficient (p-
value) Coefficient (p-value)
Intercept 29176.83 48289.42 12443.95
6.76*** 8.15*** 5.72***
INEQAV -1019.89 1266.00 -2024.22
-1.88** -1.82** -3.20***
NUMDOCS 280.67 613.58 -78.02
4.60*** 4.88*** -1.39
MONTHS -334.70 -488.67 263.06
-2.00** -1.15 1.92**
LOCATION
(GP2) -5136.82 -2139.02
-3.731** -1.95**
LOCATION (GP3) 9011.64 3923.79
2.38** 1.89**
LOCATION (GP5) 7430.43
2.09**
ANNWAGERES 12553.30 7601.19 163.29
25.28*** 8.59*** 0.19
SATPAY 1678.88
2.31**
F-Statistic 65.31*** 10.46*** 3.39***
Adj. R
2
0.54 0.50 0.15
N 613 104 153
Note: Refer to Table 2 for variable definitions.
45
possible explanations. Satisfaction with pay is added to control for the possibility
that inequity aversion proxies for low levels of satisfaction with pay. Inequity
aversion remains significantly negative while satisfaction with pay remains
significantly positive in these regressions, helping to rule out the possibility that the
two key variables are measuring the same underlying construct. Additionally, the
sample is partitioned into high and low performers to control for the possibility of
actual performance endogeneity. The results show significant negative coefficients
on inequity aversion in both cases, supporting H2.
The third hypothesis suggests that there is an interaction effect between a
specific aspect of the compensation system (subjectivity vs. objectivity) and the level
of inequity aversion on the dependent variable, performance. The first step in testing
this hypothesis is to regress performance (average monthly production) on the
inequity aversion measure, a subjective/objective dummy, and the interaction of
these two variables (along with the clinic controls). The results show a significant
coefficient on both the inequity aversion and subjective/objective dummy variables,
as well as the interaction term. This result supports the claim that an interaction
effect exists.
To further explore this effect, I used ANCOVA analysis with planned
contrasts. I classified the sample into both high and low levels (using a mean split) of
inequity aversion and subjectivity. The ANCOVA analysis again shows the
significance of the interaction effect (F-test p-value 0.00). More specifically, the
results show that when employees exhibit high levels of inequity aversion the type of
compensation system has a strong effect on their performance. When a subjective
46
Table 9
Regressions of Production on Inequity Aversion and Subjectivity/Objectivity
Interaction
This table reports the results of regressing performance, as measured by average monthly production (AVMOPROD), on the
aversion to inequity in pay (INEQAV), a categorical variable denoting subjectivity (SUBJDUM), and the interaction of these
variables (INEQAV*SUBJDUM). *** denotes significance at the 0.01 level (2-tailed). ** denotes significance at the 0.05 level
(2-tailed). * denotes significance at the 0.10 level (2-tailed).
ANNPROD
Coefficient (p-value)
Intercept 29813.20
5.75***
INEQAV -1493.03
-1.67*
SUBJDUM -1107.90
-1.73**
INEQAV*SUBJDUM -1845.00
-1.79**
MONTHS -327.39
1.58
YRSEXP 101.46
1.24
NUMDOCS -36.88
-1.23
ANNWAGERES 12530.55
0.21
GENDER 2639.02
-1.23
LOCATION (GP2) -3270.71
-2.16**
F-Statistic 58.82***
Adj. R
2
0.47
N 476
SUBJDUM = An indicator variable which equals one for all employees who receive subjective compensation. Subjective
compensation is defined as a difference of more than $5,000 between predicted wages and actual wages (see data section for
description).
INEQAV*SUBJDUM = An interaction of the inequity aversion and subjective indicator variables.
Note: Refer to Table 2 for additional variable definitions.
47
Table 10
ANCOVA Test Inequity/Subjectivity Interaction
Panel A denotes the results of ANCOVA analysis of the differences in the means of groups formed by interacting aversion to
inequity in pay (INEQAV) and subjectivity (SUBJDUM). *** denotes significance at the 0.01 level (2-tailed). ** denotes
significance at the 0.05 level (2-tailed). * denotes significance at the 0.10 level (2-tailed). Panel B denotes these results
graphically.
Panel A: Mean Comparisons
Objective
Scheme Subjective Scheme
Inequity Averse 34041.20 29366.04
4675.16
(.000)***
Self-Interested 32306.10 32001.11
304.99
(.689)
1735.10 (2635.07) F-Test
(.012)*** (.000)*** (.000)***
Panel B: Graphical Interpretation of Mean Effects
system is in place, the negative performance effect for inequity averse individuals is
exacerbated. However, when there is no preference for inequity aversion, the
performance difference between the subjective and objective systems is non-
significant. These results support H3.
48
Sensitivity Analyses
I conduct additional analyses to ensure the results presented above are not
driven by data analysis decisions. In addition to the average monthly production
indicator of performance, I also conduct all tests using an annual and single month
production measures to control for the missing data problem described above.
Results were qualitatively similar. The measure of subjectivity was also created
using several different splits including median, $1,000, and $5,000. No changes to
the definition of the variable created differences in the results. I also run the
subjectivity regression using a continuous variable of the differences between
expected and actual wages with similar results. Finally, I split high/low inequity
aversion using the median, and top/bottom 40%. I find no changes to the results
presented above.
49
Chapter 4: Conclusion
4.1 Discussion
This study highlights the claim of economists and psychologists alike that an
incentive program will be more likely to produce the desired results when it is better
matched with the preferences of its agents. These preferences have moved from the
traditional simplifying assumption of caring about only your own wealth and leisure
to include various social preferences. The results show that when the incentive
program is aligned with the preferences of the individual, both self-interested and
social, performance will increase. When these incentives are aligned with the
preferences of the individual we see the desired performance improvements from the
incentive compensation system. This result demonstrates the heterogeneity of
preferences and justifies the use of tailored incentives to increase performance within
a firm.
These additional other-regarding preferences suggested by the economics
literature are also illustrated by some responses of the surveyed medical practitioners
to open-ended questions. “Everything is about money and maximizing profit to the
detriment of working in a fun environment, practicing good medicine, and feeling
satisfied in my job.” Another medical practitioner maintains that he “would rather
have increased support staff than increased compensation,” supporting the idea that
he would forego his own monetary compensation for the good of the team (an
obvious social preference). These comments suggest that the employees of ABC
Health exhibit preferences that are clearly not solely self-interested.
50
One specific social preference which is addressed by both the theoretical and
experimental economics literatures is the preference to avoid inequity in pay.
Differences in pay between individuals in the same role, both for larger and smaller
amounts, create some disutility for the individual. This disutility may then translate
into performance problems under incentive systems. When an individual is inequity
averse his production will suffer as compared to a more self-interested other, given
the compensation system differences that exist within the firm. The results here show
that the performance of individuals who particularly dislike inequity is significantly
lower than their inequity neutral counterparts. This result provides some evidence
that firms should consider the implications of employing multiple compensation
methods within the same job level. It also suggests that some of the mixed empirical
evidence regarding the performance-incentive relationship may be due to the absence
of social preferences in their incentive models.
Not only do the empirical results above suggest the relationship between
decreased performance and aversion to differences in pay, free response comments
from the medical practitioners reported in the survey also support this claim. One
medical practitioner explained that “Although having commission-based pay is nice
it causes competition among medical practitioners which I don’t like.” Another
medical practitioner claimed, “I enjoy my practice and [ABC Health] as a whole, but
it is very disappointing and disconcerting to see other medical practitioners in other
practices provide poor-to-substandard care yet earning substantially more.” The
medical practitioners also contend that these differences affect their production. For
51
example, “There is no incentive to produce more when new graduates routinely earn
salaries much higher than mine.”
Finally, managerial accounting research has shown that subjectivity in
incentives can have positive performance effects, but can also be detrimental if it is
used incorrectly. The findings here suggest that there is some level of inequity
aversion among the employees of the firm. The paper argues that when employees
have a high level of inequity aversion they will perform even worse when they
cannot clearly attribute their (or others) compensation differences to effort, as is the
case under a subjective system. The results show that there is a significant interaction
between the type of compensation system and the level of inequity aversion. Since a
performance based incentive system automatically introduces differences in pay, by
its very nature, employers should carefully consider the addition of subjectivity in
their compensation system. For employees who dislike these differences in pay,
subjectivity can exacerbate the resulting performance problems, while objectivity
can mitigate them.
4.2 Firm Outcomes
The firm, after reading the results presented here, has put together a team to
analyze common characteristics of the medical professionals in their field. They
hope that by learning more about trends in the field (who is entering, leaving, etc.)
they will better be able to create a compensation contract which suits a wider variety
of their employees.
Management of the firm was unaware of the implications of leaving the
current compensation systems in place after the acquisition of a new unit. After my
52
analysis, I spent several days with the management of the firm detailing the findings
from both an economic and theoretical perspective. I also worked with a team from
the firm to compile an analysis on the cost of changing over the compensation
systems of the current employees to a “standard” compensation contract. While the
firm has agreed that the change is necessary, they plan to implement it gradually as
the medical practitioner’s contracts come up for review. The feedback from the field
management team, as well as the medical practitioners themselves, has been positive.
4.3 Limitations
The limitations of this paper stem first from the method employed. While the
research site presents an interesting setting to test the hypotheses, it is a single firm
within a single industry which limits its generalizability. Economic research has
shown that healthcare firms, and more specifically clinics, often behave differently
than other industries in relation to business strategy.
18
These differences, however,
do not seem to impact the effectiveness of incentives, as the health care industry
suffers from the same mixed incentive performance results as the economy as a
whole (see Armour et al., 2001; Conrad et al., 1998; Hillman et al., 1989).
Other aspects of the firm also pose a threat to my results in that the
measurements used to proxy for the underlying variables are not ideal. The
18
Specifically, incentives in the clinic environment may be theoretically different due to the
relationship between medical practitioners, clinics, and patients. The medical practitioner first has a
responsibility to the patient to provide the best care, but he has also entered into contracts with the
clinic itself (Harris, 1977). The medical practitioner feels ethically drawn to center on the needs of his
patient. However, changes in the economy have forced him to consider the costs of the treatment and
the patient’s ability to pay them. This creates a contracting dilemma: if you provide incentives to the
medical practitioner based on an overall production number he is pushed to prescribe expensive tests
(Andersen & Company, 1977). On the other hand, if you provide incentives based on efficiency he
may be pushed to provide less than the standard of care.
53
management of the firm had significant input into the survey design process.
Management feared that the survey would stir up compensation issues within the
firm and thus limited the scope of the survey.
A second limitation is that several surveys were returned stating the
employee does not know how others are compensated. This lack of knowledge of the
differences in pay, however, would bias against my finding results for hypotheses
two and three.
4.4 Contributions
This paper opens the door to studying both social preferences and the
interaction of these preferences with incentive systems. A natural extension of this
project would be to examine whether differences in the compensation system within
a single operating unit cause that unit’s performance to suffer. A second related
extension is to determine whether other measures of performance suffer the same
effects when social preferences are considered. For instance, does customer
satisfaction suffer when incentives are not well aligned, or when the employee has a
strong preference for inequity aversion?
Overall, this paper contributes to the academic literature by supporting
theoretical and laboratory evidence that social preferences do play a part in
individuals’ preference functions. The introduction of these preferences not only
affects performance, but also changes the way in which incentives systems function.
This research contributes to the field by broadening our understanding of incentive
systems by further understanding the factors which impact the effectiveness of the
incentive-performance relationship. This paper is also one of the first to examine the
54
effect of incentives on both individually measured indicators of performance and
lower level employees. Additionally, this paper contributes to practice by informing
managers that incentive programs tailored to employee preferences can be both
justified and beneficial. However, creating inequity in pay can have a negative
performance effect which is exacerbated by the introduction of subjectivity to the
compensation system.
55
Bibliography
Andersen & Company, A. 1977. Study of Reimbursement and Practice
Arrangements of Provider -Based Physicians. In Final Report to the
HealthCare Financing Administration.
Anderson, S. W., H. C. Dekker, and K. L. Sedatole. 2007. An Empirical
Examination of Negotiated Goals and Performance-to-Goal Following the
Introduction of an Incentive Bonus Plan. In Working Paper: Rice University.
Andreoni, J., M. Castillo, and R. Petrie. 2003. What do Bargainers' Preferences Look
Like? Experiments with a Convex Ultimatum Game. American Economic
Review 93:672-685.
Andreoni, J., and J. Miller. 2000. Giving According to GARP: An Experimental Test
of the Rationality of Altruism. In Mimeo: University of Wisconsin
Carnegie Mellon University.
Armour, B. S., M. M. Pitts, R. Maclean, C. Cangialose, M. Kishel, H. Imai, and J.
Etchason. 2001. The Effect of Explicit Financial Incentives on Physician
Behavior. Arch Intern Med 161 (10):1261-1266.
Ashton, R. H. 1990. Pressure and Performance in Accounting Decision Settings:
Paradoxical Effects of Incentives, Feedback, and Justification. Journal of
Accounting Research 28:148-180.
Awasthi, V., and J. Pratt. 1990. The Effects of Monetary Incentives on Effort and
Decision Performance: The Role of Cognitive Characteristics. The
Accounting Review 65 (4):797-811.
Baiman, S. 1982. Agency research in managerial accounting: A survey. Journal of
Accounting Literature 1:154-213.
———. 1990. Agency research in managerial accounting: a second look.
Accounting, Organizations and Society 15:341-371.
Bandiera, O., I. Barankay, and I. Rasul. 2005. Social Preferences and the Response
to Incentives: Evidence from Personnel Data. Quarterly Journal of
Economics 120:917-962.
56
Banker, R. D., S. Y. Lee, and G. Potter. 1996a. A Field Study of the Impact of a
Performance Based Incentive Plan. Journal of Accounting and Economics 21
(2):195-226.
Banker, R. D., S. Y. Lee, G. Potter, and D. Srinivasan. 1996b. Contextual Analysis
of Performance Inputs of Outcome-Based Incentive Compensation. The
Academy of Management Journal 39 (4):920-948.
———. 2000. An Empirical Analysis of Continuing Improvements Following the
Implementation of a Performance Based Compensation Plan. Journal of
Accounting and Economics 30 (3):315-350.
Bartling, B., and F. von Siemens. 2004. Inequity Aversion and Moral Hazard with
Multiple Agents: University of Munich.
Berg, J., J. Dickhaut, and K. McCabe. 1995. Trust, Reciprocity and Social History.
Games and Economic Behavior X:122-142.
Bol, J. C. 2007. The Determinants and Performance Effects of Supervisor Bias. In
Working Paper: The University of Illinois at Urbana-Champaign. Working
Paper.
Bonner, S. E., R. Hastie, G.B. Sprinkle, & S.M. Young. 2000. A review of the
effects of financial incentives on performance in laboratory tasks:
implications for management accounting. Journal of Management
Accounting Research 12:19-64.
Bouwens, J., and L. van Lent. 2006. Performance measure properties and the effect
of incentive contracts. Journal of Management Accounting Research 18:55-
75.
Brandts, J., and G. Charness. 2004. Do Labour Market Conditions Affect Gift
Exchange? Some Experiemental Evidence. Economic Journal 114:684-708.
Burks, S., J. P. Carpenter, and L. Goette. 2006. Performance Pay and the Erosion of
Worker Cooperation: Field Experimental Evidence. In Discussion Paper
Series #2013: Institute for the Study of Labor.
Cameron, L. 1999. Raising the Stakes in the Ultimatum Game: Experimental
Evidence from Indonesia. Economic Inquiry 37 (1):47-59.
Carpenter, J. P., and E. Seki. 2006. Competitive work environments and social
preferences: Field experimental evidence from a Japanese fishing
community. BE Journal in Economic Analysis and Policy: Advances 5 (2).
57
Charness, G. 2004a. Attribution and Reciprocity in an Experimental Labor Market.
Journal of Labor Economics 22 (July):665-688.
———. 2004b. Attribution and Reciprocity in an Experimental Labor Market.
Journal of Labor Economics 22 (3):665-688.
Charness, G., and M. Rabin. 2002. Understanding Social Preferences with Simple
Tests. Quarterly Journal of Economics (August):817-869.
Cole, B. S. 1986. Cole Compensation Survey -- Average Base Salary Increase Slips a
Notch to 6.7% in 1986/More Facilities Use Incentives to Reward Execs.
Modern Healthcare 16 (23):65.
Conrad, D. A., C. Maynard, A. Cheadle, S. Ramsey, M. Marcus-Smith, H. Kirz, C.
A. Madden, D. Martin, E. B. Perrin, T. Wickizer, B. Zierler, A. Ross, J.
Noren, and S. Liang. 1998. Primary Care Physician Compensation Method in
Medical Groups: Does it Influence the Use and Cost of Health Services for
Enrollees in Managed Care Organizations? Journal of the American Medical
Association 279 (11):853-858.
Core, J., and W. R. Guay. 2001. Stock Option Plans for Non-Executive Employees.
Journal of Financial Economics 61 (2):253-287.
Dawes, R., and R. Thaler. 1988. Cooperation. Journal of Economic Perspectives 2
(3):187-197.
Demski, J. S., and G. A. Feltham. 1976. Cost determination: A conceptual approach.
Ames, IA: Iowa State University Press.
———. 1978. Economic Incentives in Budgetary Control Systems. The Accounting
Review 53:336-359.
Denis, D., P. Hanouna, and A. Sarin. 2006. Is There a Dark Side to Incentive
Compensation? Journal of Corporate Finance 12:467-488.
Eisenhardt, K. M. 1989. Agency Theory: An Assessment And Review. Academy of
Management. The Academy of Management Review 14 (1):57.
Englmaier, F., and A. Wambach. 2005. Optimal Incentive Contracts under Inequity
Aversion. In SSRN Working Paper: Harvard Business School.
Evans, J. H., R. L. Hannan, R. Krishnan, and D. V. Moser. 2001. Honesty in
Managerial Reporting. The Accounting Review 76 (4):537-559.
58
Falk, A., E. Fehr, and U. Fischbacher. 2000. Informal Sanctions: Institute for
Empirical Research in Economics, University of Zurich.
Fehr, E., and A. Falk. 2002. Psychological Foundations of Incentives. European
Economic Review 46:687-724.
Fehr, E., and U. Fischbacher. 2002. Why Social Preferences Matter - The Impact of
Non-selfish Motives on Competition, Cooperation, and Incentives. The
Economic Journal 112 (March):C1-C33.
Fehr, E., G. Kirchsteiger, and A. Riedl. 1993. Gift Exchange and Ultimatum in
Experimental Markets. Vienna Economic Papers: University of Vienna
Department of Economics vie 9301.
Fehr, E., and K. M. Schmidt. 1999. A Theory of Fairness, Competition, and Co-
operation. Quarterly Journal of Economics 114:817-868.
Fehr, E., and E. Tougareva. 1995. Do High Monetary Stakes Remove Reciprocal
Fairness? Experimental Evidence from Russia. In MIMCO: Institute for
Empirical Economic Research, University of Zurich.
———. 2002. Do High Stakes and Competition Undermine Fairness? Evidence
from Russia. In SSRN Working Paper.
Fernie, S., and D. Metcalf. 1996. It's Not What You Pay It's The Way You Pay It
And That's What Gets Results: Jockeys Pay and Performance. London School
of Economics Discussion Paper 295.
Fessler, N. J. 2003. Experimental Evidence on the Links among Monetary
Incentives, Task Attractiveness, and Task Performance. Journal of
Management Accounting Research 15:161.
Forsythe, R., J. Horowitz, N. E. Savin, and M. Sefton. 1994. Fairness in Simple
Bargaining Experiments. Games and Economic Behavior 6 (3):347-369.
Franco, F., C. Ittner, and D. Larcker. 2007. The Effects of Contract Design and
Implementation Practices on Worker Incentive Plan Outcomes: University of
Pennsylvania.
Gibbs, M., K. A. Merchant, W. A. Van der Stede, and M. E. Vargus. 2004.
Determinants and Effects of Subjectivity in Incentives. The Accounting
Review 79 (2):409-436.
59
Groves, T., Y. Hong, J. McMillan, and B. Naughton. 1994. Autonomy and
Incentives in Chinese State Enterprises. The Quarterly Journal of Economics
109 (1):183-209.
Grund, M., and D. Sliwka. 2002. Compassion and Envy in Tournaments. IZA
Discussion Paper 647.
Guth, W., R. Schmittberger, and B. Schwarze. 1982. An Experimental Analysis of
Ultimatum Bargaining: an Experimental Study. Journal of Economic
Behavior and Organization 3:367-388.
Hannan, R. L. 2000. The effect of firm profit on fairness perceptions, wages and
employee effort. Ph.D., University of Pittsburgh, United States --
Pennsylvania.
Hannan, R. L., J. Kagel, and D. V. Moser. 2002. Partial Gift Exchange in an
Experimental Labor Market: Impact of Subject Population Differences,
Productivity Differences, and Effort Requests on Behavior. Journal of Labor
Economics 20 (4):923-951.
Hannan, R. L., F. W. Rankin, and K. L. Towry. 2006. The Effect of Information
Systems on Honesty in Managerial Reporting: A Behavioral Perspective.
Contemporary Accounting Research 23 (4):885-918.
Harris, J. E. 1977. The internal organization of hospitals: some economic
implications. The Bell Journal of Economics 8 (2):467-482.
Heijden, E. C. M., J. H. M. van der Nelissen, J. J. M. Potters, and H. A. A. Verbon.
2001. Simple and Complex Gift Exchange in the Laboratory. Economic
Inquiry 39 (2):280-297.
Heneman, H. G. I., and D. P. Schwab. 1985. Pay satisfaction: Its multidimensional
nature and measurement. International Journal of Psychology 20:129-141.
Hillman, A. L., M. V. Pauly, and J. J. Kerstein. 1989. How do financial incentives
affect physicians' clinical decisions and the financial performance of health
maintenance organizations? N Engl J Med 321 (2):86-92.
Hogarth, R. M., B. J. Gibbs, R. M. McKenzie, and M. A. Marquis. 1991. Learning
from Feedback: Exactingness and Incentives. Journal of Experimental
Psychology: Learning, Memory, and Cognition 17:734-752.
Indjejikian, R. J. 1999. Performance Evaluation and Compensation Research: An
Agency Perspective. Accounting Horizons 13 (2):147-157.
60
Jenkins, G. D. 1986. Financial Incentives. In Generalizing from laboratory to field
settings, edited by E. A. Locke. Lexington, MA: D.C. Hearth and Company,
167-180.
Jenkins, G. D., A. Mitra, N. Gupta, and J. D. Shaw. 1998. Are financial incentives
related to performance? A meta-analytic review of empirical research.
Journal of Applied Psychology 83:777-787.
Jensen, M. C., and W. H. Meckling. 1976. Theory of the firm: Managerial behavior,
agency costs and ownership structure. Journal of Financial Economics 3
(4):305-360.
Judge, T. A., and T. M. Welbourne. 1994. A confirmatory investigation of the
dimensionality of the pay satisfaction questionnaire. Journal of Applied
Psychology 79:461-466.
Jutras, C. 2007. Executive Summary Benchmark Report 2007 [cited October 18
2007].
Kagel, J., and K. Wolfe. 2001. Tests of fairness models based on equity
considerations in a three-person ultimatum games. Experimental Economics
4:203-219.
Kahneman, D., and W. S. Peavler. 1969. Incentive Effects and Pupillary Changes in
Association Learning. Journal of Experimental Psychology 79:312-318.
Kirchsteiger, G. 1994. The Role of Envy in Ultimatum Games. Journal of Economic
Behavior and Organization 25:373-389.
Lawler, E. E. I. 1971. Pay and organizational effectiveness. New York: McGraw-
Hill.
———. 1973. Motivation in the work place. Monterey, CA: Brooks/Cole.
Lazear, E. P. 2000. Performance Pay and Productivity. The American Economic
Review 90 (5):1346-1361.
Ledford, G. E., E. E. I. Lawler, and S. A. Mohrman. 1995. Reward innovations in
Fortune 1000 companies. Compensation and Benefits Review 27:76-80.
Ledyard, J. 1995. Public Goods: A Survey of Experimental Research. In Handbook
of Experimental Economics, edited by J. Kagel and A. E. Roth. Princeton:
Princeton University.
Lee, T. W., E. A. Locke, and S. H. Phan. 1997. Explaining the Assigned Goal-
Incentive Interaction: The Role of Self-Efficacy and Personal Goals. Journal
of Management 23:541-559.
61
Libby, R., and M. G. Lipe. 1992. Incentives, Effort, and the Cognitive Processes
Involved in Accounting-Related Judgments. Journal of Accounting Research
30 (2):249-273.
Luft, J. L. 1997. Fairness, Ethics, and the Effect of Management Accounting on
Transaction Costs. Journal of Management Accounting Research 9:199-216.
Mas, A. 2006. Pay, reference points, and police performance. Quarterly Journal of
Economics 121 (3):783-821.
Matsumura, E. M., and T. Y. Shin. 2006. An Empirical Analysis of a Relative
Performance Based Incentive Plan: Evidence from a Postal Service. The
Accounting Review 81 (3):533-566.
McPheron, T. 2007. 2007 is DVM Year of the Woman 2007 [cited October 18 2007].
Milgrom, P., and J. Roberts. 1992. Economics, Organization, and Management.
Englewood Cliffs, NJ: Prentice Hall.
Minkler, L. 2004. Shirking and Motivations in Firms: Survey Evidence on Worker
Attitudes. International Journal of Industrial Organization 22:863-884.
Neilson, W. S., and J. Stowe. forthcoming. Piece-Rate Contracts for Other-
Regarding Workers. Economic Inquiry.
Opsahl, R. L., and M. D. Dunnette. 1966. The role of financial compensation on
industrial motivation. Psychological Bulletin 66:94-118.
Ostrom, E., J. Walker, and R. Gardner. 1992. Covenants with and without a Sword:
Self Governance is Possible. American Political Science Review 86:404-417.
Paarsch, H., and B. Shearer. 1999. The Response of Worker Effort to Piece-Rates:
Evidence from the British Columbia Tree-Planting Industry. Journal of
Human Resources 34:643-667.
Pfeffer, J., and N. Langton. 1993. The effect of wage dispersion on satisfaction,
productivity, and working collaboratively: Evidence from college and
university faculty. Administrative Science Quarterly 38 (3):382-407.
Pfeffer, J., and R. I. Sutton. 2006. What's wrong with pay-for-performance.
Industrial Management 48 (2):12-17.
Prendergast, C. 1999. The Provision of Incentives in Firms. Journal of Economic
Literature 37:7-63.
62
Rabin, M. 1993. Incorporating Fairness into Game Theory and Economics. The
American Economic Review 83 (5):1281-1302.
———. 2002. A Perspective on Psychology and Economics. European Economic
Review 46:657-685.
Rotemberg, J. 1994. Human Relations in the Workplace. Journal of Political
Economy 102 (4):684-717.
Roth, A. E., M. W. K. Malouf, and J. K. Murningham. 1981. Sociological versus
Strategic Factors in Bargaining. Journal of Economic Behavior and
Organization 2:153-177.
Shah, J. E., T. Higgins, and R. S. Friedman. 1998. Performance incentives and
means: how regulatory focus influences goal attainment. Journal of
Personality and Social Psychology 74:285-293.
Sprinkle, G. B. 2000. The Effect of Incentive Contracts on Learning and
Performance. The Accounting Review 75 (3):299-326.
Tuttle, B., and F. G. Burton. 1999. The effects of a modest incentive on information
overload in an investment analysis task. Accounting, Organizations and
Society 24 (8):673-687.
U.S. Bureau of Economic Analysis, G. 2006. National income and product account
tables 2006 [cited April 16, 2007 2006]. Available from
www.bea.gov/bea/dn/nipaweb.
Vroom, V. H. 1964. Work and motivation. New York: Wiley.
Weiner, B. 1966. Motivation and Memory. Psychological Monographs: General and
Applied 80:1-22.
Wright, P. M. 1989. Test of the Moderating Role of Goals in the Incentive-
Performance Relationship. Journal of Applied Psychology 74:699-705.
Wright, P. M., and K. M. Kacmar. 1995. Moderating Roles of Self-Set Goals, Goal
Committment, Self-Efficacy, and Attractiveness in the Incentive-
Performance Relationship. Human Performance 8:263-296.
Abstract (if available)
Abstract
This study examines whether the inclusion of social preferences in the assumed utility function of individuals can provide a richer understanding of the incentive-performance relationship. Recent economic research has shown that individuals' utility functions consist of not only their own wealth and leisure, but also social preferences such as reciprocity, honesty, and inequity aversion. This paper tests whether incentive systems, all employed within a single firm, have differing effects on performance when their fit with several aspects of utility is considered. The results show that an increased fit between incentive systems and individual preferences is associated with increased performance. Additionally, when employees exhibit high levels of a specific social preference -- aversion to differences in pay -- performance suffers under a pay-for-performance scheme. Finally, when an employee exhibits high levels of aversion to differences in pay and is under a subjective incentive scheme, the negative performance effects are exacerbated. This research contributes to the field by broadening our understanding of incentive systems by further understanding the factors which impact the effectiveness of the incentive-performance relationship.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
The interactive effects of incentive threshold and narcissism on managerial decision-making
PDF
The effects of accounting performance and professional relationships on promotion, dismissal, and transfer decisions in a conglomerate
PDF
Utility functions induced by certain and uncertain incentive schemes
PDF
Marketing strategies with superior information on consumer preferences
PDF
Essays on the empirics of risk and time preferences in Indonesia
PDF
Essays on understanding consumer contribution behaviors in the context of crowdfunding
PDF
Dynamic social structuring in cellular self-organizing systems
PDF
Essays on delegated portfolio management under market imperfections
PDF
Reward shaping and social learning in self- organizing systems through multi-agent reinforcement learning
PDF
Experience modulates neural activity during action understanding: exploring sensorimotor and social cognitive interactions
PDF
A social-cognitive approach to modeling design thinking styles
PDF
Essays on information design for online retailers and social networks
PDF
Essays on service systems with matching
PDF
Designing infectious disease models for local level policymakers
PDF
Three essays on cooperation, social interactions, and religion
PDF
Modeling social and cognitive aspects of user behavior in social media
PDF
Architectural innovations for mitigating data movement cost on graphics processing units and storage systems
PDF
Understanding the psycho-social and cultural factors that influence the experience of attention-deficit/hyperactivity disorder (ADHD) in Chinese American college students: a systems approach
PDF
Essays on improving human interactions with humans, algorithms, and technologies for better healthcare outcomes
PDF
Crowding in peripheral vision
Asset Metadata
Creator
Martin, Melissa Anne
(author)
Core Title
Incorporating social preferences into incentive system design
School
Marshall School of Business
Degree
Doctor of Philosophy
Degree Program
Business Administration
Degree Conferral Date
2008-12
Publication Date
12/10/2008
Defense Date
08/08/2008
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
compensation,incentives,inequity aversion,OAI-PMH Harvest,social preferences
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Merchant, Kenneth A. (
committee chair
), Kim, Peter H. (
committee member
), Van der Stede, Wim A. (
committee member
)
Creator Email
martinm@usc.edu,Melissa.Martin@asu.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-m1906
Unique identifier
UC179759
Identifier
etd-Martin-2350 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-141793 (legacy record id),usctheses-m1906 (legacy record id)
Legacy Identifier
etd-Martin-2350.pdf
Dmrecord
141793
Document Type
Dissertation
Rights
Martin, Melissa
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Repository Name
Libraries, University of Southern California
Repository Location
Los Angeles, California
Repository Email
cisadmin@lib.usc.edu
Tags
compensation
incentives
inequity aversion
social preferences