American Economic Association
Reference Points and Effort Provision
Author(s): Johannes Abeler, Armin Falk, Lorenz Goette and David Huffman
Source: The American Economic Review, Vol. 101, No. 2 (APRIL 2011), pp. 470-492
Published by: American Economic Association
Stable URL: http://www.jstor.org/stable/29783680
Accessed: 18-05-2015 12:21 UTC
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact [email protected].
American Economic Association is collaborating with JSTOR to digitize, preserve and extend access to The American Economic
Review.
http://www.jstor.org
This content downloaded from 131.220.108.137 on Mon, 18 May 2015 12:21:21 UTC
All use subject to JSTOR Terms and Conditions
American Economic Review 101 (April 2011): 470-492
101.2.470
http://www.aeaweb. org/articles.php ?doi=10.1257/aer.
Reference Points and EffortProvision
By Johannes
Abeler,
Armin
Falk,
Lorenz
Goette,
and David
Huffman
A key open question for theories of reference-dependent preferences
is expecta?
is: what determines the reference point? One candidate
tions: what people expect could affect how theyfeel about what actu?
the rational
ally occurs. In a real-effort experiment, we manipulate
influ?
expectations of subjects and check whether this manipulation
ences their effort provision. We find that effortprovision
is signifi?
cantly different between treatments in theway predicted by models of
if expectations
reference-dependent preferences:
expectation-based,
are high, subjects work longer and earn more money than ifexpecta?
tionsare low. (JELD12, D84, J22)
Imagine two identical workers. One expected a salary increase of ten percent but
receives an increase of only five percent. The other receives the same five percent
wage increase but had not expected an increase. The change in income is the same
presumably feels less satisfied. Intuitively,
outcomes
what
in
of
many people judge
they expected to happen. In this paper,
light
we test this particular notion: whether expectations serve as a reference point.
and
A growing class of theories (e.g., David E. Bell
1985; Graham Loomes
for both workers,
but the first worker
Robert Sugden 1986; Faruk Gul 1991; JonathanShalev 2000; Botond Koszegi
is built on the idea that expectations can
Rabin 2006, 2007, 2009)
are
as
a
act
able to align empirical evidence that
reference point. These models
and Sugden
is hard to reconcile with usual economic assumptions
(e.g., Loomes
and Koszegi
2008; Fabian Herweg, Daniel M?ller, and Philipp
1987; Paul Heidhues
and Matthew
their theoretical and intuitive appeal, models of expec?
tation-based, reference-dependent
preferences are inherently difficult to test, as
expectations are hard to observe in the field. To sidestep this problem, we conduct a
tightly controlled, real-effort experiment. The two main advantages of our setup are
Weinschenk
2010).
Despite
thatwe know the rational expectations
of participants
in regard to earnings and that
*Abeler:
University of Nottingham, School of Economics, Nottingham, NG7 2RD, UK (e-mail: johannes.
Falk: University of Bonn, Adenauerallee
24,53113 Bonn, Germany (e-mail: armin.falk@
[email protected]);
Internef, 1015 Lausanne, Switzerland
uni-bonn.de); Goette: University of Lausanne, Department of Economics,
Swarthmore College,
Economics
Huffman:
500 College
Department,
(e-mail: [email protected]);
Financial
Swarthmore, PA 19081 (e-mail: [email protected]).
support from the Deutsche
thank Steffen Altmann, Roland
15 is gratefully acknowledged. We
through SFB/TR
Forschungsgemeinschaft
Lucie D?rner, Simon G?chter, Uri Gneezy, Paul Heidhues,
Oliver Kirchkamp, Botond Koszegi,
Benabou,
Dorothea K?bler, Juanjuan Meng, Konrad Mierendorff, Susanne Ohlendorf, Andrew Oswald, Matthew Rabin, Sam
Schulhofer-Wohl, Klaas Schulze, Barry Schwartz, Daniel Seidmann, Chris Starmer, Kostas Tatsiramos, Christine
and Matthias Wibral for helpful discussions. Valuable comments were also received
Valente, Georg Weizs?cker,
from numerous seminar and conference participants. Anke Becker, Franziska Tausch, Benedikt Vogt, and Ulf Z?litz
Ave.,
provided excellent research assistance.
470
This content downloaded from 131.220.108.137 on Mon, 18 May 2015 12:21:21 UTC
All use subject to JSTOR Terms and Conditions
VOL.
101 NO.
2
ABELER
ETAL.:
REFERENCE
POINTSAND
EFFORT
PROVISION
471
we can exogenously
influence these expectations. We are thus able to directly assess
the relevance of theories of expectation-based,
reference-dependent preferences.
Investigating the importance of expectations helps with answering the key open
for reference-dependent preferences: what determines the reference point?
an empirically validated theory of where reference points come from
Developing
question
is crucial for disciplining predictions. Otherwise, if the reference point is assumed
case-by-case, models of reference-dependent preferences might explain behavior not
because
of their structural assumptions but because of this additional degree of free?
dom. Testing expectations as potential candidate for a reference point extends previous
empirical research which has restricted attention mainly to the status quo or lagged
status quo as reference point (e.g., Daniel Kahneman,
Jack L. Knetsch,
and Richard H.
Thaler 1990;TerranceOdean 1998;David Genesove andChristopher
Mayer 2001).
In our experiment, subjects work on a tedious and repetitive task. After each rep?
etition, they decide whether to continue or to stop working. They get a piece rate,
but receive their accumulated piece rate earnings only with 50 percent probability,
whereas with 50 percent probability they receive a fixed, known payment instead.
Which
their
payment subjects receive is determined only after they have made
a
choice about when to stop working. The only treatment manipulation
is variation
in the amount of the fixed payment.
the amount of the fixed payment, we change the effort incen?
By manipulating
to the models
tives of subjects with reference points in expectations. According
mentioned
above, these individuals experience painful loss sensations if the realized
state of theworld compares unfavorably to a state that they expected could possibly
happen instead. In our experiment, a subject can expect one of two states of the
to occur, with equal probability, when they stop working: receiving the fixed
payment,/, or receiving piece rate earnings. Anticipating potential disappointment
if they were to receive the less favorable of the two states, subjects can minimize
potential loss sensations by stopping with piece rate earnings close to/. This way,
no matter what happens, they know that unfavorable comparisons will not be too
world
severe.1 From
of minimizing
losses, the best plan is actually to stop
a subject expects / in either state of
case
to
in
this
with earnings exactly equal
/;
the world, outcomes necessarily fulfill these expectations, and there is no chance
the perspective
of disappointment whatsoever. But whether it is optimal to completely eliminate
potential loss, or instead just stop closer to/ depends on an individual's effort costs.
Our treatmentmanipulation
thus exogenously raises earnings expec?
increases/and
tations for the case
that the fixed payment is realized. This means that an individual
has to choose a higher effort level to reduce or eliminate a potential feeling of loss if
they do not receive/. On the other hand, they still do not want towork too far beyond
/ to avoid being verydisappointed in case theydo receive/Assuming a reference
point in expectations thus predicts that increasing the size of the fixed payment will
tend to increase overall effort, and that the propensity to stop is especially high when
the piece
rate earnings equal
the fixed payment.
1
Note thatwhile choosing to stop with piece rate earnings unequal to/would allow for the possibility of a gain,
if themore favorable outcome is realized, gain sensations are typically weaker than loss sensations (Kahneman and
Amos Tversky 1979). Due to this asymmetry, known as "loss aversion," the subject would ratherminimize devia?
tions from expectations.
This content downloaded from 131.220.108.137 on Mon, 18 May 2015 12:21:21 UTC
All use subject to JSTOR Terms and Conditions
APRIL 2011
THEAMERICANECONOMIC REVIEW
472
By contrast, a canonical model of effortprovision with separable utility over money
and effort costs does not predict a treatment difference. Optimal effort is determined
by settingmarginal cost equal to themarginal benefit defined by the piece rate, and
the fixed payment is irrelevant for both marginal cost and marginal benefit. This is true
independent of the shape of utility over money and the shape of the cost function, con?
ditional on the assumption of separability. Models
incorporating reference-dependent
preferences, but taking the status quo as the reference point, also predict no treatment
difference, because the status quo is the same across treatments.
Our data support themain predictions of reference-dependent preferences models
the amount of the fixed payment is
with a reference point in expectations. When
large, subjects work significantly more
is small. We also observe pronounced
the amount of the fixed payment
spikes in the distribution of effort choices,
amount
at
in the low fixed payment treatment, and at
low
the
fixed
payment
exactly
there
the high fixed payment amount in the high fixed payment treatment.Moreover,
is no spike at the high fixed payment amount in the low fixed payment treatment,
than when
and vice versa. In additional control treatments, we show that our results are not
driven by alternative, psychological mechanisms:
subjects do not stop exactly at the
fixed payment because this number is especially salient, and they do not work more
when the fixed payment is higher because of reciprocal feelings toward the experi?
menter. Finally, we provide evidence that reference-dependent preferences are the
the degree of
behind our results from another angle: We measure
key mechanism
subject by having them make a series of small-stakes lottery
choices after the experiment. We find that subjects who are more loss averse accord?
stop closer to the fixed payment. This is a unique
ing to this independent measure
a
of
the
reference
of
theory
point in expectations, where stronger loss
prediction
loss aversion of each
aversion
One
leads to greater attraction to/.
specific application of our findings
is to the literature on labor supply and
series of studies have found evidence consistent with
transitory wage changes. A
loss aversion around a daily reference income (e.g., Colin Camerer et al. 1997; Yuan
K. Chou 2002; Ernst Fehr and Lorenz Goette 2007; Henry S. F?rber 2008; Vincent
Crawford and Juanjuan Meng
forthcoming; Kirk Bennett Doran 2009), with the
exception of F?rber (2005). In this literature the reference point has typically been
treated as an unobserved,
latent variable. Most closely related to our paper is the
recentstudyby CrawfordandMeng (forthcoming)
who use data onNew York City
taxi drivers' labor supply to test the theoryof Koszegi and Rabin (2006). They
proxy the rational expectation about a driver's wage by the average wage earned per
weekday and find evidence for income and hours targeting around this expectation.
there is no experimental variation, they address the problem of endogene
Because
ity using a structural approach. Our approach is complementary, in using a tightly
laboratory setting that allows us to exogenously vary rational expecta?
tions regarding earnings. Our studies find converging evidence on the importance of
reference points in expectation for effort provision. We discuss the implications for
the labor supply literature inmore detail in Section V.
controlled
Also
related to our paper
in which
in lottery choices,
is the literature on violations
of expected utility theory
some findings are supportive of a role for expectation
based referencepoints (seeLoomes and Sugden 1987; SyngjooChoi et al. 2007; and
Andreas Hack
and Frauke Lammers
2008 for discussions).
Different from our paper,
This content downloaded from 131.220.108.137 on Mon, 18 May 2015 12:21:21 UTC
All use subject to JSTOR Terms and Conditions
VOL.
101 NO.
2
ABELER
ETAL.:
REFERENCE
POINTSAND
EFFORT
PROVISION
473
has mainly come from inconsistencies observed in choices involv?
ing relatively complex combinations of different financial lotteries. Our experiment
adds to this literature by measuring
the impact of reference points as expectations
this evidence
in the domain
of real-effort choices, rather than lottery decisions. Moreover,
it pro?
vides corroborating evidence on the importance of reference points as expectations,
based on a simple and transparent test, where subjects can act in accordance with
expectedutilitytheorysimplyby ignoringthefixedpayment.
The paper is organized as follows. Details of the experimental design are explained
in the following section. Section II discusses behavioral predictions. Results of the
two main treatments are presented in Section III. Section IV reports on the control
treatments and the lottery-based measure of loss aversion. Section V concludes.
I. Design
Our experiment was designed to create an environment that allows a precise mea?
surement of behavior and inwhich we can exogenously
influence a reference point
in expectations. In the experiment, subjects worked on a tedious task. As the work
task, we chose counting the number of zeros in tables that consisted of 150 ran?
domly ordered zeros and ones. This task does not require any prior knowledge,
and there is little learning possibility; at the same
performance is easily measurable,
time, the task is boring and pointless, and we can thus be confident that the task
entailed a positive cost of effort for subjects. The task was also clearly artificial, and
output was of no intrinsic value to the experimenter. This minimizes
any tendency
for subjects to use effort in the experiment as a way to reciprocate for payments
offered by the experimenter.
experiment involved two stages. Prior to the first stage, subjects read the
instructions and answered control questions; theywere also told that the experiment
had a second stage but that details would be provided later.2During the first stage,
subjects had fourminutes to count as many tables as possible. They received a piece
rate of ten cents per correct answer for sure.3 This part served to familiarize subjects
with the task; due to this first stage, subjects had a good understanding of how dif?
The
ficult the task was
one could earn in a given time before they knew
the amount of the fixed payment (which was revealed only after the first
stage).
Additionally, we will use performance in this stage as a productivity indicator.
After the first stage, subjects read the instructions for the second
(and main) stage.
The task was again to count zeros, but there were two differences compared to the
first stage. First, they could now decide themselves how much and for how long
they wanted to work. At most, they could work for 60 minutes. When
they wanted
to stop, they could push a button on the screen and the experiment was over: sub?
and how much
jects answered a very short questionnaire
(including a series of small-stakes lottery
choices described in Section IV), got paid immediately, and could leave. How much
subjects chose to work will be the main outcome variable in our analysis of the
2
An English translation of the instructions is provided in theweb Appendix.
3
In both stages, if an answer was not correct, subjects had two more tries for the same table. To prevent guess?
ing, the piece rate was deducted from their account if they failed all three tries.This happened only 59 times in the
experiment (compared to almost 12,000 correctly counted tables).
This content downloaded from 131.220.108.137 on Mon, 18 May 2015 12:21:21 UTC
All use subject to JSTOR Terms and Conditions
474
THEAMERICANECONOMIC REVIEW
APRIL 2011
experiment. The second difference was that subjects did not get their accumulated
piece rate earnings from themain stage for sure. Before they started counting in the
main stage, they had to choose one of two closed envelopes. They knew that one of
contained a card saying "Acquired earnings" and that the other enve?
a card saying "3 euros." But they did not know which card was in
contained
lope
which envelope. The envelopes remained with the subjects while theywere working
and were only opened after the subject had stopped working. The subject's payment
the envelopes
was
then determined by the card in the chosen envelope. The piece rate per correct
doubled to 20 cents in themain stage in order to keep economic incen?
answer was
tives comparable between the two stages.
We know the rational expectation of each subject regarding earnings, in the
moment they were deciding whether to stop working: with 50 percent probability
the subject would
he would receive
receive the accumulated piece rate earnings and with 50 percent
the fixed payment. Because
uncertainty about the payment was
was
we
were able to exogenously vary a sub?
revealed only after thework
finished,
amount
the
rational
of the fixed payment.4
ject's
expectation by changing
There were
two main treatments. The only difference between these treatments
the amount of the fixed payment. In the LO treatment, the fixed payment was
three euros while itwas seven euros in theHI treatment. Treatments were assigned
was
randomly to subjects; we also randomized treatments over morning and afternoon
time slots and over days of the week.
A potential confound could have arisen if subjects worked in the same room and
started working, e.g., due to peer effects (Falk and Andrea Ichino
simultaneously
2006) or due toa desire forconformity(B. Douglas Bernheim 1994).We employed
a special procedure to prevent such effects: subjects arrived for the experiment
one at a time, and individual starting times were at least 20 minutes apart. Upon
arrival, subjects were guided to one of three essentially identical, neutral rooms.5
alone in their room with the door closed and never (with very few
saw
another subject or the other two experimental rooms. Instructions
exceptions)
and payments were also administered in their room. Because
of this special proce?
They worked
dure, subjects' stopping behavior could not have been influenced by other subjects'
behavior in a systematic way.
We conducted three additional control treatments to check whether salience or
reciprocity could have driven the results. Design
described in Section IV.
and results of these treatments are
Subjects were students from the University of Bonn studying various majors
except Economics. We recruited subjects who had participated in no or only a few
previous experiments. Experiments were computerized using the software z-Tree
and ORSEE
in each
(Urs Fischbacher 2007, Ben Greiner 2004). 60 subjectsparticipated
subject participated inmore than one treatment. In addition
from the two stages of the experiment (on average 8.70 euros),
treatment. No
to their earnings
4We don't know the actual expectations of subjects. The theories we are testing, however, all rely on the theoreti?
cal construct of rational expectations. Even so, it is unlikely that actual expectations are far from correct formany
subjects in our setting: the lotterywas very simple and salient, and the potential payoffs (current accumulated piece
rates and fixed payment) were always shown on the screen.
5
Photos of the three rooms are shown in theweb Appendix.
This content downloaded from 131.220.108.137 on Mon, 18 May 2015 12:21:21 UTC
All use subject to JSTOR Terms and Conditions
VOL.
101 NO.
2
ABELER
ETAL.:
REFERENCE
POINTSAND
EFFORT
subjects received a show-up fee of five euros. The experiment
on average, including time for instructions and both stages.
PROVISION
475
took about one hour
II. Predictions
three categories of models: a canonical model with separable utility,
and models with expectation-based
with status-quo reference dependence,
can
be described as follows: The subject's choice
reference dependence. Our setup
We
examine
models
is the number of correctly solved tables e.With probability 1/2 each, the
receives
either a fixed payment / or the accumulated piece rate earnings
subject
w
>
0
is the piece rate per table. c(e) is the subject's cost of effortwith
where
we,
we are
dc/de > 0 that the subject has to bear in both states of the world. Because
variable
interested in the effect of the size of/on
and HI, respectively.6
effort provision, we
set/to fLO and/^
for
treatments LO
First, consider a standard model of effort provision with a utility function sepa?
?
rable inmonetary payoff x and cost of effort: U(x, e) = u(x)
c(e). In our setup,
thisutilityfunctionbecomes U(e,f, w) = u(f)/2 + u(we)/2 ? c(e), yielding the
following first-order condition:
if the subject
The optimal effort level e* is independent of the fixed payment/;
no
matter
to
how large/is.
receives the fixed payment/ he wishes
stop right away
The prediction depends on the separability of money and cost of effort7,but not on
the shape of the cost function nor on a subject's curvature in w(-), i.e., whether the
subject is risk-neutral, risk-averse, or risk-loving.
Models
incorporating reference-dependent preferences with the status quo as the
the status quo when
reference point also predict no treatment difference, because
across
same
treatments
thus
is
and
the
the
independent of/. The
entering
experiment
can
to
status
rate
the
how
affect
quo, but effort can?
earnings compare
subject
piece
loss relative to status quo if the fixed payment is realized.
Thus, varying the amount of the fixed payment has no impact on effort incentives,
although the individual has reference-dependent preferences. A similar argument
not influence the potential
holds for reference points thatmay be affected by expectations about earnings that
subjects had before learning about the exact incentive scheme for their particular
treatment.
6
As
it happened only very rarely that a subject miscounted a table thrice and thus got the piece rate deducted
from his earnings, we ignore this design detail in the predictions.
7It is common in labor economics to allow for nonseparability across time periods in effort cost (e.g., fatigue) or
in consumption utility (e.g., habit formation). With such forms of nonseparability, themodel still predicts no treat?
ment difference. Adopting a less common assumption of contemporaneous nonseparability of income and effort,
?
=
e.g., with a function U
c(e)], allows generating the prediction that effort increases with/, if
g[u((we + f)/2)
the function g[-] is concave, i.e., if increasing expected wealth makes counting zeros less painful. The equally plau?
sible assumption that having more money makes counting zeros more painful predicts that effort should decrease
with /. Either way, such a utility function cannot explain the tendency to stop exactly at the fixed payment nor a
correlation between individual loss aversion and stopping closer to/. Adding further assumptions to capture these
other predictions becomes more and more ad hoc and specific to our experimental setup and makes it impossible to
generalize themodel to other settings. By contrast, the formulation of the reference-dependent preferences model
below is inspired by a large body of empirical evidence and is easily generalized to different settings.
This content downloaded from 131.220.108.137 on Mon, 18 May 2015 12:21:21 UTC
All use subject to JSTOR Terms and Conditions
476
the american economic review
april 2011
In contrast to these two models,
theories assuming that agents have expectation
preferences predict different behavior across treat?
based, reference-dependent
ments. Here, individuals dislike an outcome
falling short of their expectations. We
deriveourhypothesesusing themodel ofKoszegi andRabin (2007); models byBell
(1985), Loomes and Sugden (1986), andGul (1991) generatesimilarpredictions.8
IV, we derive further hypotheses for three control
lottery-based measure of individual loss aversion.
In Section
treatments and the
an individual derives "consumption utility" from
In Koszegi
and Rabin
(2007),
the consumption bundle c and "gain-loss utility" from comparing c to a reference
bundle r. Bundle r is the full distribution of rational expectations,
i.e., every out?
come
that could have happened weighted with its ex-ante probability. As outcomes
in our setup are not very large, we assume consumption utility to be linear and
equal to c. Overall utility is the sum of consumption and gain-loss utility, and is
assumed to be separable across the K dimensions of c. We assume that subjects
assess
outcomes along two dimensions: money and effort costs. The gain-loss util?
?
and Rabin
is
defined
by the function fi(ck
ity
rk). For small arguments s, Koszegi
=
=
>
s
assume that
is
linear:
for
0
and
rjs
piece-wise
rjXs for s < 0
^(s)
fi(s)
fj,(s)
>
with tj
0 and A > 1; because A is strictly greater than 1, losses loom larger than
equal-sized gains.9 We assume that the gain or loss sensation a subject finally experi?
ences depends on their rational expectations about possible earnings amounts held
themoment before the envelope is opened. The final piece rate earnings (and the
fixed payment amount for the treatment) thus determine the reference point. This
corresponds to the choice-acclimating
(2007), where the individual's choice
concept
equilibrium
in Koszegi
and Rabin
and thus the reference
shapes expectations,
point that is held right before uncertainty is resolved. Because
subjects pay the effort
costs regardless of which envelope they draw, effort costs do not affect the compari?
son between the two states of the world and thus do not enter gain-loss utility.10
If the subject intends to stop at an accumulated
earnings
level below
ment (we < /), theresultingexpectedutilitywill be given by
U=
^-
c(e) + 1?
L(f-we)
y(we
-
the fixed pay?
we) + ~X(we
f)
+ !(/-/)
The first two terms are expected consumption utility and cost of effort.The remain?
ing terms are the expected gain-loss utility: the first bracketed term is the gain-loss
utilitywhen theoutcome iswe,multipliedby theprobabilityof occurring(1/2) and
8The main difference between Koszegi and Rabin (2007) and the other theories is how expectations are mapped
into a reference point. Bell (1985), for example, assumes it to be themean while Koszegi and Rabin (2007) assume
that an outcome is compared to the entire distribution of expectations; this distinction does not matter for our setup.
9
In its full generality, themodel assumes that a stochastic outcome F is evaluated according to its expected util?
ity,with the utility of each outcome being the average of how it feels relative to each possible realization of the ref?
=
erencepointG: U(F G)
belieftheindividual
pointG is theprobabilistic
\
f J u(c \r)dG(r)dF(c).The reference
held in the recent past about outcomes.
10Effortcosts do still enter consumption utility.As for the canonical model, we make the simplifying assumption
that subjects know their effort costs. One reason for the first stage of the experiment was to give subjects experience
with their individual effort costs.
This content downloaded from 131.220.108.137 on Mon, 18 May 2015 12:21:21 UTC
All use subject to JSTOR Terms and Conditions
VOL.
101 NO.
2
ABELER
ETAL.:
REFERENCE
POINTSAND
EFFORT
477
PROVISION
by 77,the strength of gain-loss utility. Inside this term, expecting and receiving we
feels neutral; but receiving we while expecting the larger/feels like a loss. Since the
subject expected to receive/with
probability 1/2, the terms are weighted accord?
ingly. The second bracketed term shows gain-loss utility where the outcome is the
fixed payment, applying the same logic.
>
If the accumulated earnings are higher than the fixed payment
(we
/), the gain
now
loss utility is different. Receiving
the accumulated
feels like a gain
earnings
compared to the lower fixed payment (third term), while receiving
ment now means a loss (terms equal to zero are suppressed
here):
U =
The
we
?^
-
first-order conditions
When
c(e)+ 1?
1
(we
-
f)
the fixed pay?
Mf
we)
are then:
earnings are below/, themarginal returns to effort are higher
is the return to effort in the canonical model without gain-loss
accumulated
than w/2, which
utility(assuming linearw( )). Stopping entails a loss if theoutcome turnsout to
be we rather than /;
the pain of this loss more than offsets the potential pleasure
of a gain iff is realized. When
the accumulated earnings are above/
the incentive
effect of loss aversion is reversed: because earnings beyond/can
be lost in case the
loss aversion now reduces the returns to effort
subject receives the fixed payment/
relative to the canonical case. Gain-loss
utility thus creates an additional incentive
to exert effort when below the fixed payment amount, and reduces the incentive
to work when
above the fixed payment. Therefore, increasing the fixed payment
should increase average effort, since it causes themarginal returns to remain high
up to a higher effort level.
1: Average
HYPOTHESIS
effort in the HI
treatment is higher
than in the LO
treatment.
Reference dependence moves optimal effort from above and below towards the
fixed payment; themore loss averse a subject is the closer they should stop to the
fixed payment. For some subjects, this will even move optimal effort so far as to
equalize expected piece rate earnings and/. The discrete drop in the return to effort
at the fixed payment amount implies that there is a range of cost functions forwhich
stopping exactly at the fixed payment is optimal. Thus, there will tend to be cluster?
ing of stopping decisions exactly at/ The stronger loss aversion
the larger the fraction stopping at/will be.
HYPOTHESIS
than in theHI
is in the population,
2: The probability to stop at we = fL0 is higher in the LO treatment
treatment; theprobability to stop at we = fHI is higher inHI than inLO.
This content downloaded from 131.220.108.137 on Mon, 18 May 2015 12:21:21 UTC
All use subject to JSTOR Terms and Conditions
478
We
THEAMERICANECONOMIC REVIEW
next turn to the empirical
APRIL 2011
results from the experiment.
III. Results
Our first result supports Hypothesis
1. In the LO treatment with fixed payment
? 3
euros, subjects stop working after accumulating 7.37 euros on average. In the
/
HI treatment with/ = 7 euros, subjects stop on average at 9.22 euros.
RESULT
the LO
1: Subjects
treatment.
in theHI
treatment work significantly more
than subjects
in
treatment difference of 1.85 euros is almost half as large as the amount of
?
the treatment manipulation
3 ? 4 euros). The marginal effect compared
to
(7
effort provision in LO is 25.1 percent. The treatment difference in effort provision
The
is significant in an OLS
regression where we compare effort inHI to effort in LO.
the
accumulated
regress
earnings at which a subject stopped on a treatment
dummy (see Table 1, column l).11 The treatment difference stays significant when
we control for productivity, gender, outside temperature
(experiments took place in
the summer), and time of day. The only significant control variable is productivity
We
(Table 1, columns 2 and 3). As an indicator for productivity in themain stage, we
use average time per correct answer in the first stage
(measured in seconds mul?
?1
A
coefficient
thus
indicates
that
faster subjects complete
tiplied by
positive
).
more
tables.12
It could be that the cost of effort is not only determined by the number of tables
counted but also by themere time subjects spend in the experiment. We therefore con?
sider the time spent working as an alternative measure of effortprovision. Treatments
are also different for this dependent variable: subjects in LO work on average 31.7
minutes, while subjects inHI work on average 6.4 minutes longer, a marginal effect
of 20.1 percent. This difference is significant inOLS regressions with and without the
controls described above (see Table 1, columns 4 - 6).13 Because
subjects can only
work between 0 and 60 minutes, we also present Tobit regressions that account for
this censoring
(Table
1, columns
7-9).
This does not alter the results.14
As shown inSection II, themodel ofKoszegi andRabin (2007) predictsthatstop?
ping decisions in the two treatments should differ in a very special way. Hypothesis 2
predicts a higher probability of stopping when the accumulated
earnings equal the
11
The result is confirmed by nonparametric tests.A Mann-Whitney
t/-testyields a p-value of 0.015 (all p-values
in this paper refer to two-sided tests). The same result obtains ifwe compare the distribution of stopping decisions:
=
a two-sample Kolmogorov-Smirnov
test rejects the equality of distributions between treatments
0.005).
(p
12The Spearman rank correlation coefficient between answering speeds in each stage is 0.520 (p < 0.001).
This measure of productivity is not influenced by the treatmentmanipulation
since subjects during the first stage
did not know yet about the exact procedure of themain stage. Consequently, answering speed in the first stage is
=
not significantly different between treatments
({/-test, p
0.185). Using average time per answer (i.e., including
also wrong answers) or number of completed tables during the first stage instead of themeasure used above does
not change results.
13The treatment difference in working time is also statistically significant in nonparametric tests: {/-test,
= 0.034 ;
test,p = 0.085.
p
Kolmogorov-Smirnov
an
not
we
is
issue
if
take earnings as dependent variable; earnings are neither bounded above nor
14Censoring
below (since subjects could make losses by miscounting tables thrice).
This content downloaded from 131.220.108.137 on Mon, 18 May 2015 12:21:21 UTC
All use subject to JSTOR Terms and Conditions
VOL.
101 NO.
2
ABELER
Table
1?Treatment
ETAL.:
REFERENCE
Difference
OLS: Accumulated earnings
1 ifHI treatment
in Effort
EFFORT
PROVISION
toLO
treatment)
(HI compared
OLS: Time spentworking
(inmin.)
479
Tobit: Time spentworking
(inmin.)
(1)
(2)
(3)
(4)
(5)
(6)
(8)
(9)
1.850**
1.942**
1.973**
6.430**
6.572**
6.784**
7.927*
8.091**
8.442**
(3.163)
(3.153)
0.091
(3.231)
0.096
(3.841)
(3.814)
0.098
(3.833)
0.103
(0.067)
(0.070)
1.619
(0.080)
(0.083)
1.577
(0.917)
Productivity
(0.885)
(0.900)
0.059*** 0.064***
(0.019)
1 ifFemale
Controls for
POINTSAND
No
No
No
No
(0.020)
-0.039
(0.950)
Yes
(7)
No
No
(3.412)
Yes
No
No
(4.035)
Yes
No
No
Yes
No
No
Yes
temperature
Controls fortime
of day
Constant
Observations
Adjusted or Pseudo R2
Yes
7.370***
10.607*** 10.200***
(0.648)
(1.206) (1.445)
31.715*** 36.713*** 34.362**
(2.237)
(4.297) (5.190)
33.004*** 38.389*** 35.306**
(2.697)
(5.143) (6.116)
120
120
120
120
120
120
120
120
120
0.03
0.09
0.08
0.03
0.03
0.00
0.00
0.01
0.01
Notes: The dependent variable is the level of accumulated earnings (in euro) atwhich a subject stoppedworking for columns 1-3,
and time spentworking (inminutes) until a subject stopped for columns 4-9. Columns 1-6 reportresultsfromOLS regressions,
columns 7-9 show resultsof Tobit regressions (the lower and upper limitsare 0 and 60 minutes). Data fromLO and HI treatments
are included in theanalysis. The proxy forproductivityis the time subjects needed per table during the firststage (in secondsmul?
?
tipliedby 1). Standard errorsare inparentheses.Adjusted R2 is shown forOLS; pseudo R2 forTobit.
***Significant at the 1 percent level.
**
Significantat the5 percent level.
*
Significantat the 10 percent level.
fixed payment. Neither the canonical model, nor themodel with status quo reference
dependence, make this prediction. Our data are consistent with Hypothesis 2.
to stop when accumulated
2: The probability
earnings are equal to the
amount of thefixed payment is higher compared to the same earnings level in the
other treatment. The modal choice in both treatments is to stop exactly when accu?
mulated earnings equal thefixed payment.
RESULT
Figure 1 shows a histogram of accumulated earnings (LO in the top panel, HI in
the bottom panel). First of all, stopping decisions are dispersed over a wide range.
Some subjects stop directly, others work for up to 25 euros. This is what one would
expect given that productivity and cost of effort differ across subjects. But there
are systematic differences between treatments in terms of
clustering of stopping
decisions exactly at the fixed payment: in the LO treatment, many subjects stop at
three euros
(15.0 percent of subjects); in the HI treatment, almost nobody stops at
three euros (1.7 percent). By contrast, inHI many subjects stop at seven euros
(16.7
percent); inLO veryfew subjectsstophere (3.3 percent).The modal choice inboth
treatments is to stop exactly when accumulated earnings equal the fixed payment.
treatment differences are statistically significant. Results of a multinomial
logit regression with the three outcomes "stop at 3 euros," "stop at 7 euros," and
1 shows the regression without
"stop elsewhere" are presented in Table 2. Column
These
controls, in columns 2 and 3 the controls used in Table 1 are added. Being in the
HI treatment leads to significantly less stopping at three euros and more stopping at
This content downloaded from 131.220.108.137 on Mon, 18 May 2015 12:21:21 UTC
All use subject to JSTOR Terms and Conditions
APRIL 2011
THEAMERICANECONOMIC REVIEW
480
LO
15
10
5_
c
0
o
cd
Q
0
IM I i iiIL
ilUiu?
0.
1
11
13
15
17
19
21
23
25
Hl
15
10
5J
0?
I m iIii11
0
1
3
Figure
5
7
11
ilVp
Accumulated
earnings
9
13
of Accumulated
1.Histogram
at Which
a Subject
15
17
Earnings
Stopped
i in
19
21
,?r
23
25
(in Euros)
seven euros compared to being in the LO treatment.15The same results obtain ifwe
compare the number of subjects stopping in a range around three and seven euros.
For example between two and four euros, 30.0 and 5.0 percent of subjects stop in
and HI, respectively (Latest, p < 0.001); between six and eight euros, these fig?
ures are 13.3 and 38.3 percent, respectively ([/-test, p = 0.002). Multinomial
logit
estimates for this result are presented in Table A.l in theWeb Appendix.
LO
IV. Robustness
Checks
In the previous section we presented evidence supporting models of expectation
based, reference-dependent preferences: subjects work more when expectations are
high, and many subjects stop when piece rate earnings equal the fixed payment, thus
avoiding any potential loss relative to expectations.
In this section, we check whether other, psychological
motivations
could
also
have played a role in generating the observed treatment differences. We first pres?
ent results of three control treatments showing that neither salience nor reciprocity
in our setting. We
then provide further,
significantly influences effort provision
a
direct evidence that loss aversion is key mechanism
driving our findings: We use
an independent measure of an individual's degree of loss aversion,
more loss averse subjects stop closer to the fixed payment.
and show that
15These differences are also significant in nonparametric tests: the percentages of subjects stopping at three
euros is significantly higher in LO ({/-test, p = 0.009);
the percentage stopping at seven euros is higher in HI
=
(?/-test,p
0.015).
This content downloaded from 131.220.108.137 on Mon, 18 May 2015 12:21:21 UTC
All use subject to JSTOR Terms and Conditions
VOL.
101 NO.
2
Table
ABELER
ETAL.:
treatment
POINTSAND
to Stop at the Fixed Payment
2?Tendency
Stop at 3
1 ifHI
REFERENCE
Stop at 7
(la)
(lb)
1.609**
-2.197**
(1.073)
(0.801)
Stop at 3
(2a)
Stop at 7
(2b)
1.620**
-2.191**
(1.074)
(0.802)
0.005
(0.014)
(0.016)
1 ifFemale
treatment)
Stop at 3
(3b)
(3a)
-2.318**
(1.115)
-0.003
for time of day
Constant
Stop at 7
1.781**
(0.829)
0.004
(0.019)
(0.020)
(0.789)
(0.661)
0.106
-1.094
Controls for temperature
481
PROVISION
(HI compared toLO
0.003
Productivity
Controls
EFFORT
NoYes No
NoYes No
-1.695***
(0.363)
-3.199***
(0.721)
-1.523*
-2.946***
(0.848)
(1.121)
-1.437
(1.215)
-3.032**
(1.326)
120
120
Observations
0.09
Pseudo/?2
0.09
0.17
Notes: The table reports results of multinomial logit regressions. The dependent variable indicates three outcomes:
"stop at 3 euros," "stop at 7 euros," and "stop elsewhere" which is the reference category. Data from LO and HI
treatments are included in the analysis. The proxy for productivity is the time subjects needed per table during the
?
first stage (in seconds multiplied by
1). Standard errors are in parentheses.
***Significant at the 1 percent level.
**Significant at the 5 percent level.
*
Significant at the 10 percent level.
A. Salience
that stopping decisions did not cluster at the fixed payment
of reference dependence but because the fixed payment was salient. If sub?
jects resorted to irrelevant, environmental cues to decide when to stop, theymight
have stopped at three euros or seven euros because these amounts were mentioned
It is conceivable
because
frequently in the instructions and also on the computer screens and could have
served as a focal point.16 Similarly, it could be that the salience of the fixed payment
influenced subjects to set arbitrary goals of earning these amounts (Edwin A. Locke
and Gary P. Latham 1990, 2002) and thus impacted behavior through this channel
rather than reference-dependent preferences.17
To check whether the salience of, e.g., "3 euros" influences behavior
in our
experiment, we conducted two symmetric treatments, both variants of the LO treat?
treatment (NOSAL
ment. The NOSAL
for no salience)
suppresses the salience of
"3 euros"
as much
at three. The SAL
as possible but keeps the reference-dependence motive to stop
treatment (SAL for salience) keeps the salience of the fixed pay?
ment exactly the same as in the LO treatment but removes the loss aversion motive
to stop there. The general procedure of the two treatments was identical to the LO
16Focal points or arbitrary anchors have been shown to influence behavior by, e.g., Tversky and Kahneman
(1974), Karen E. Jacowitz and Kahneman
(1995), Gretchen Chapman and Eric J. Johnson (1999); and David K.
Whynes, Zoe Philips, and Emma Frew (2005).
17
In a recent paper, Alexander K. Koch and Julia Nafziger (2009) demonstrate that this intuition could also go
in the other direction. They build upon themodel of Koszegi
and Rabin (2006) to show that goal setting can be
reference-dependent preferences (see also Chip
explained by assuming that individuals have expectation-based,
Heath, Richard Larrick, and George Wu 1999).
This content downloaded from 131.220.108.137 on Mon, 18 May 2015 12:21:21 UTC
All use subject to JSTOR Terms and Conditions
482
THEAMERICANECONOMIC REVIEW
APRIL 2011
treatment: subjects came one-by-one, they counted zeros in tables after choosing
one of two envelopes, and the card in their chosen envelope determined their payoff
when they decided to stop working. We again collected 60 observations per treat?
ment. The only difference was the two cards. Recall
"Acquired earnings" and "3 euros."
that in LO,
the two cards read
In theNOSAL
treatment, the cards read "5 euros plus acquired earnings" and "8
euros." Thus, unlike in LO, the number "3" was never mentioned
in the description
or
on
screens
of the potential payoffs
the computer
and could therefore not have
acted as a convenient anchor. Still, the way to avoid potential losses was to stop at
three euros. The
five euros were taken out of the show-up fee, so total
were
to the LO treatment; the two treatments differed
identical
expected earnings
in
the
merely
framing of payoffs. If salience is important in our setting, behavior
inNOSAL
should differ from behavior in LO: there should be no special tendency
to stop at three inNOSAL.
Subjects should rather stop more often at five or eight
additional
euros than subjects inLO. As the salient numbers inNOSAL
point to larger amounts
to
LO, average effort should be higher than in LO. Expectation-based,
compared
and LO
reference-dependent preferences, however, predict that behavior inNOSAL
shouldnotbe different
and thatalso inNOSAL, subjectsshouldbe especially likely
to stop exactly at three.18
In the SAL treatment, the two cards read "Acquired earnings" and "Acquired
earnings plus 3 euros." This means that subjects in SAL actually received the accu?
mulated piece rate for sure and played an additional lottery (0, 3 euros; 0.5). To
keep incentives for a rational, risk-neutral subject the same as in LO, the piece rate
in SAL was halved to ten cents (since subjects received the piece rate only with 50
percent probability inLO but got it for sure in SAL). Salience of "3 euros" remained
exactly as in the LO treatment: every occurrence of "3 euros" in the original instruc?
tions or screens was replaced by the phrase "acquired earnings plus 3 euros" where
applicable. "3 euros" was thusmentioned equally often and at the same places as in
the LO
treatment.
If the high probability of stopping at three is driven by salience, there should be
a tendency to stop right at three in SAL, as was the case in LO. But if subjects have
expectation-based,
reference-dependent
preferences, the treatments should differ,
with no clustering of stopping decisions at three in SAL. Expected utility in SAL,
according to themodel byKoszegi andRabin (2007), is givenby:
U =
we +
f
-
-^
, v
i
c{e) + ?ri
2 J
If the subject receives fin addition to the piece rate earnings, this feels like a
gain relative to the alternative of only getting the piece rate earnings (third term).
If the subject only receives the piece rate earnings, this feels like a loss relative to
also gettingtheadditional amount/(fourthterm).The loss of/is weightedmore
heavily
than the gain of/
leading
to a lump-sum
reduction
in utility, but this does
18
The canonical model of effortprovision and a model of loss aversion around the status quo similarly predict no
difference between NOSAL
and LO. But at the same time, these models do not predict that there should be cluster?
ing of stopping decisions at three in either treatment.
This content downloaded from 131.220.108.137 on Mon, 18 May 2015 12:21:21 UTC
All use subject to JSTOR Terms and Conditions
VOL.
101 NO.
ABELERETAL.: REFERENCEPOINTSAND EFFORT PROVISION
2
483
not influence optimal effort. In contrast to the main treatments, subjects in SAL
cannot influence the size of a potential loss by choosing a particular effort level.
loss aversion does not
Therefore, unlike in the LO treatment, expectation-based
=
a
to
for
cluster
this does not
3, because
at/
tendency
predict
stopping decisions
=
/ should thus be lower in SAL
help avoid losses. The probability to stop at we
than in LO.19
primarily to test the different predictions of reference-depen?
dent preferences and salience regarding when subjects will, and will not, cluster
stopping decisions exactly at three euros. One can also compare average effort lev?
We
designed
SAL
themodels make more similar predictions. The
model
predicts that effort in SAL will be higher
reference-dependent preferences
in
most
As
in
than LO.
shown
subjects inLO stopped above the fixed pay?
Figure 1,
ment. For these subjects, loss aversion held back their effort. In SAL, the loss aver?
els in SAL
and LO,
but in this case
to hold back effort is removed, as there is no risk of feeling a loss when
the
getting
piece rates, so subjects should work harder, potentially the full time.
not reference dependence?drives
If salience?and
behavior, then there should be
sion motive
clustering of stopping decisions at three in SAL, which requires subjects to count
more tables than in LO and thus raises average effort (the nominal piece rate was
reduced to ten cents as subjects got the piece rate for sure).20
from the two control treatments do not support the salience
clustering of stopping decisions at the fixed payment amount:
Data
explanation
for
3: The probability to stop exactly at three euros is not significantly differ?
ent between NOSAL
and LO, despite the difference in salience; and subjects stop
less often at three euros in SAL compared toLO, although salience is held constant
RESULT
In the NOSAL
treatment, 13.3 percent of subjects stop at three euros compared
in the LO treatment. In SAL, only 3.3 percent of subjects stop at
to 15.0 percent
three euros. Results
of multinomial
logit regressions comparing stopping behavior
toLO are shown inTable 3.21The dependent variable iswhether
to
subjects stopped at three euros, at seven euros, or somewhere else. Compared
in
the LO treatment, stopping behavior is not different in NOSAL.
SAL,
Subjects
however, stop significantly less often at three euros. These results continue to hold
when we include the control variables used in Tables
1 and 2 (columns 2 and 3)
or ifwe consider stopping in a range around three euros; see Table A.2 in theWeb
do not stop more often at five or eight
Appendix.22 In addition, subjects inNOSAL
inNOSAL
LO
=
/
and SAL
19
The standard model of effortprovision with a separable, linear utility function predicts no difference between
?
and SAL. In such a model, SAL implies exactly the same incentives as LO: U = (we +/)/2
c(e), with
no
treatment difference.
3. Reference dependence around the status quo also predicts
20The canonical model with linear utility predicts the same effort level in SAL and LO. But ifutility is concave,
a sure piece rate of ten cents will be valued higher than a 200 piece rate with 50 percent probability. In that case,
effort in SAL should again be higher.
21
In addition, the sample contains observations from our reciprocity treatment (R treatment) which we discuss
inmore detail below. The results also hold in separate regressions where we restrict the sample to only LO and
or to LO and SAL.
NOSAL
22
The SAL-LO difference of stopping at three euros is also significant in nonparametric tests (?/-test,p = 0.027)
while the NOSAL-LO
difference is not significant ((/-test, p = 0.794). A potential concern could arise because
to
in
SAL
had
subjects
complete 30 tables to reach the fixed payment of three euros while subjects in LO needed
only 15 tables. If effort costs simply made it impossible to reach 30 tables in SAL, thiswould mechanically prevent
This content downloaded from 131.220.108.137 on Mon, 18 May 2015 12:21:21 UTC
All use subject to JSTOR Terms and Conditions
Table
3?Tendency
to Stop at the Fixed
Stop at 7
Stop at 3
1 if SAL
treatment
-1.638**
(lb)
(la)
-0.134
(0.806)
1 ifNOSAL
treatment
0.958
-0.076
(0.527)
1 ifR treatment
(1.019)
(0.861)
-0.078
-0.666
(0.592)
(1.019)
Productivity
Payment
toLO)
(Control Treatments Compared
Stop at 3
(2a)
-1.619**
(0.807)
-0.081
(0.528)
-0.584
(0.599)
0.008
(0.010)
1 ifFemale
2011
APRIL
484 THEAMERICANECONOMIC REVIEW
Stop at 7
Stop at 3
(2b)
(3b)
(3a)
-0.259
-1.989**
-0.117
(1.020)
(0.902)
0.954
(0.591)
(0.688)
0.009
(0.010)
(1.027)
0.007
(0.013)
-0.411
(0.657)
NoYesNo
for temperature
NoYesNo
Controls for time of day
Constant
(1.088)
0.007
(0.014)
-0.533
(0.455)
Controls
(0.931)
0.368
-0.605
-0.009
(1.147)
1.341
-0.052
(0.862)
Stop at 7
-1.695***
(0.363)
Observations
-3.199***
(0.721)
Pseudo/?2
-1.258**
-1.004
-2.838***
-3.076***
(0.756)
(0.620)
(0.996)
240
240240
0.04 0.04 0.07
(1.193)
Notes: The table reports results of multinomial logit regressions. The dependent variable indicates three outcomes:
"stop at 3 euros," "stop at 7 euros," and "stop elsewhere" which is the reference category. Data from LO, SAL,
and R treatments are included in the analysis. The proxy for productivity is the time subjects needed per
NOSAL,
? 1
table during the first stage (in seconds multiplied by
). Standard errors are in parentheses.
***Significant at the 1 percent level.
**Significant at the 5 percent level.
*
Significant at the 10 percent level.
euros than subjects in LO, showing that just mentioning a number saliently does not
create a tendency to stop there in our setting.23Moreover,
stop
subjects inNOSAL
more often at three euros than subjects in HI ([/-test, p = 0.016) while this is not
true for SAL ([/-test, p = 0.560). These results suggest that subjects in the main
treatments do not stop at/because
of salience. In contrast, the results are consistent
reference dependence: People stop at
with themodel assuming expectations-based
three when this helps minimize
losses, even if three isminimally salient, and do not
exhibit a special tendency to stop at three if stopping there cannot help avoid losses,
although three is salient.
Data
on average effort levels are also in line with predictions
reference-dependent
RESULT
preferences:
is no difference between NOSAL
and LO
in SAL work significantly more than subjects
4: There
sion. Subjects
of expectation-based,
in average
inLO.
effortprovi?
and Tobit estimates of average effort regressed on treatment
dummies without and with the controls described above. Accumulated
earnings when
Table
4 shows OLS
differentbetweenNOSAL
stoppingare not significantly
and LO
(columns 1-3)
subjects from stopping exactly at three euros. However, 65 percent of subjects in SAL completed at least 30 tables
but only the above mentioned 3.3 percent stopped exactly at 30.
?
and 3.3 percent of LO-subjects
231.7 percent of NOSAL-subjects
stop at eight euros (CAtest: p
0.560). In both
treatments, 10 percent of subjects stop at five euros.
This content downloaded from 131.220.108.137 on Mon, 18 May 2015 12:21:21 UTC
All use subject to JSTOR Terms and Conditions
VOL.
101 NO.
Table
Difference
A?Treatment
in Effort
OLS: Accumulated earnings
(1)
(Control treatments compared
(4)
(5)
(6)
(?)
12.417*** 12.878*** 15.429***
(4.468)
-6.017
(4.424)
-6.118
(4.778)
-5.955
-6.816
(4.467)
-6.923
(4.772)
-6.541
(0.857)
-1.173
(0.856)
-1.004
(0.942)
-1.269
(4.468)
-5.933
(4.420)
-4.432
(4.841)
-5.809
(4.536)
-6.330
(4.477)
-4.780
(4.841)
-5.995
(0.857)
(0.867)
0.014
(0.983)
0.014
(4.468)
(0.011)
(0.011)
-0.311
1 ifFemale
No
No
No
No
(4.461)
0.123**
(0.050)
(0.709)
Yes
(4.522)
(9)
15.403**
-1.119
Productivity
Controls for
(8)
12.417*** 12.897**
-1.195
treatment-1.183
1 ifR treatment
toLO)
Tobit: Tables completed
OLS: Tables completed
(3)
(2)
1 ifSAL treatment
1 ifNOSAL
485
ABELERETAL.: REFERENCEPOINTSAND EFFORT PROVISION
2
(5.000)
0.130**
(4.529)
(4.511)
0.128**
(0.050)
(0.050)
-2.264
(5.005)
0.136**
(0.050)
-2.137
No
No
(3.145)
Yes
No
No
(3.148)
Yes
No
No
Yes
No
No
Yes
temperature
Controls fortime
of day
Constant
Observations
Adjusted or Pseudo R2
7.370**
8.135*
(0.606)
(0.869)
Yes
:
9.455**
(1.090)
37.050*** 43.825*** 51.045**
(3.159)
(4.155) (4.989)
37.050*** 44.100*** 51.633***
(3.198)
(4.199) (4.992)
180
180
180
240
240
240
240
240
240
0.00
0.01
0.01
0.08
0.10
0.11
0.01
0.01
0.02
Notes: The dependent variable is the level of accumulated earnings (in euro) at which a subject stoppedworking forcolumns 1-3,
and thenumber of tables completed correctlyfor columns 4-9. Columns 1-6 reportresults fromOLS regressions,columns 7-9
show resultsof Tobit regressions (the lower limit is 0 tables). Data fromLO, SAL, NOSAL, and R treatmentsare included in the
analysis (SAL only forcolumns 4-9). The proxy forproductivityis the time subjects needed per tableduring thefirststage (in sec?
ondsmultiplied by ? 1). Standard errorsare inparentheses.Adjusted R2 is shown forOLS; pseudo R2 forTobit.
***Significant at the 1percent level.
**
Significantat the5 percent level.
*
Significantat the 10 percent level.
although salient numbers (five and eight) are higher than the salient number three in
to compare earnings
LO. As the nominal piece rate in SAL differs, it ismisleading
between SAL and LO. We thus take the number of correctly completed tables as
measure of effort in columns 4-6. Subjects in SAL work significantly more than in
LO; this is consistent with loss aversion motives having held back effort in LO. This
also holds in Tobit regressions that account for the fact that the number of correctly
solved tables cannot be negative (columns 7-9).24 Taking time worked as alterna?
tivemeasure
work
of effort yields a similar result; interestingly, subjects inNOSAL
=
in
work
shorter
while
SAL
subjects
significantly longer
slightly
(f/-test,p
0.095)
than in LO (t/-test,p = 0.001). 30.0 percent of subjects in SAL work even the full
60 minutes compared to 13.3 percent in LO and 10.0 percent inNOSAL.
Finally,
the treatment difference in effort between HI and NOSAL
accumulated
earnings:
?/-test,/?<
is highly significant (e.g.,
0.001).
=
difference is also significant in nonparametric tests (tables completed:
24The SAL-LO
tZ-test,p
0.005;
=
difference is not significant (accumulated earnings:
test,p
Kolmogorov-Smirnov
0.018) while theNOSAL-LO
?
=
=
f/-test,p
test, p
0.305; Kolmogorov
0.432; tables completed:
t/-test,p
0.310; Kolmogorov-Smirnov
?
Smirnov test,p
0.432). The results also hold in separate OLS regressions where we restrict the sample to only
or to LO and SAL.
LO and NOSAL
This content downloaded from 131.220.108.137 on Mon, 18 May 2015 12:21:21 UTC
All use subject to JSTOR Terms and Conditions
486
THEAMERICANECONOMIC REVIEW
APRIL 2011
B. Reciprocity
potential motive, distinct from salience, could be reciprocity (Rabin
studies have shown thatmany indi?
Falk
and Fischbacher
1993;
2006). Numerous
even
not
acts
if
in
it is
their immediate interest (for an overview
viduals reward kind
Another
see Fehr and Simon G?chter 2000). Itmight thusbe thatsubjects inHI worked
they perceived the higher average pay inHI as a kind act
and reciprocated by counting more tables. Note, however, that
reciprocity (unlike reference-dependent preferences) cannot explain why many sub?
jects stop exactly at the fixed payment; reciprocity could potentially explain only the
more
than in LO
because
by the experimenter
result on average effort.
While
reciprocity has been
shown to be important inmany contexts, it is likely
to be irrelevant in our setting for several reasons.25 Nevertheless, we conducted an
additional control treatment to testwhether the level of total pay triggers a recipro?
cal reaction. The R treatment (R for reciprocity) is identical to the LO treatment
except for an additional lump sum payment of four euros which is announced at
the beginning of the main stage and is clearly linked to working on the task (the
instructions read "For the work in this part of the experiment, you receive 4 euros
up front."). This raises total expected earnings relative to the LO treatment and even
surpasses expected earnings in the HI treatment.26 If the higher effort in HI com?
pared to LO was simply due to a reciprocal reaction to a lump-sum payment of
seven rather than three, then average effort should also be higher inR than inLO. If,
however, reference-dependent preferences drove the treatment difference, behavior
inR should not be different from LO.
RESULT
5: Effort provision
inR and LO
is not significantly different.
by
regressions where we regress average effort measured
on treatment dummies
or
tables
earnings
completed
(LO treatment is the omitted
category) and various controls described above. We find that average effort is not
Table
4 shows OLS
significantly different inR compared to LO.
a reduction in effort.27Table 3 additionally
If anything, the point estimates suggest
shows that clustering of stopping deci?
the treatment dif?
sions in R is also not significantly different from LO. Moreover,
is
ference in effort between HI and R
highly significant (e.g., accumulated earnings:
25
First, previous studies show a role for reciprocity primarily between subjects, rather than between subjects
and the experimenter where social distance is probably larger (George A. Akerlof 1997; Gary Charness, Ernan
Haruvy, and Doron Sonsino 2007); second, direct tests of whether subjects feel reciprocal towards the experimenter
conclude that this is in fact not the case (Bj?rn Frank 1998); third, counting zeros clearly has no intrinsic value
to the experimenter and thus working hard does not benefit the experimenter; fourth, there is suggestive evidence
rather than reciprocity (Stephen Burks, Jeffrey
that piece rate compensation tends to cue income maximization,
in the instructions that subjects were free to work as
Carpenter, and Goette 2009); fifth,we strongly emphasized
little or as much as theywanted.
26
We chose an additional payment of four euros tomake sure that, even if subjects disregard the 50 percent
probability of getting the fixed payment inHI and just focus on the "nominal" earnings, perceived earnings in R
are still as high as inHI.
27The result holds also in nonparametric tests (accumulated earnings: ?/-test,/?= 0.178; Kolmogorov-Smirnov,
?
?
0.587; tables completed: Latest, p = 0.180; Kolmogorov-Smirnov,
p
p
0.587) or ifwe take time worked as
=
=
variable
0.250; Kolmogorov-Smirnov,/?
dependent
([/-test,/?
0.432).
This content downloaded from 131.220.108.137 on Mon, 18 May 2015 12:21:21 UTC
All use subject to JSTOR Terms and Conditions
VOL.
101 NO.
2
ABELER
ETAL.:
REFERENCE
U-test, p < 0.001).28 We therefore conclude
ence between the LO and theHI treatment.
C. Evidence
POINTSAND
EFFORT
487
PROVISION
that reciprocity did not drive the differ?
on Individual Loss Aversion
and Stopping Decisions
After each of our experiments, subjects answered a short questionnaire. Among
other questions, we asked subjects to state reasons for their stopping decision.
Answers were given in free form without any suggestion of possible reasons. Of
those subjects stopping exactly when accumulated
earnings equal the fixed pay?
if
the great majority named reasons such as a desire to avoid disappointment
or
to
to
the
the
favorable
that
drew
less
wanted
"make
sure"
get
envelope,
they
they
amount of the fixed payment by working at least thatmuch. Because
they indicated
ment,
a preference to avoid unfavorable comparisons
to what might have happened, we
answers
as
that
these
evidence
reference dependence and disap?
interpret
providing
pointment aversion were
payment.
But itwould
important for generating clustering of stopping at the fixed
be more desirable
to have an incentivized measure of subjects' indi?
to directly investigate the link between strength of loss aversion
on the one hand and stopping behavior on the other. During the questionnaire, we
vidual
loss aversion
subjects make six choices, each time between a fixed payment of
zero and a small-stakes lottery.The lottery involved a
50/50 chance of winning six
euros or receiving Y. Across lotteries, Y was varied from ?2 to ?7 euros in steps of
1 euro. Subjects knew that one of the six choices would be randomly selected and,
therefore had
the lottery, this lotterywould be played out formoney. Note that
the small stakes mean that rejections of lotteries with positive expected value cannot
be explained by standard risk aversion (Rabin 2000). Rather, the number of lotter?
if they had chosen
ies that a subject rejects gives an indicator for the individual's degree of loss aver?
sion. A very similar measure has been used in previous studies and has been shown
to predict loss-averse behavior in terms of labor supply, as well as strength of the
endowment effect (Fehr and Goette 2007; G?chter, Andreas Herrmann, and Johnson
2007). We thus take the number of rejected lotteries as a proxy for the individual's
degree of loss aversion. For our setting, it does not matter whether this proxy con?
cerns the strength of the gain-loss utility, rj, or the loss aversion parameter, A.29
If subjects have no reference-dependent preferences, we should not expect sub?
jects to reject small-stakes lotteries with positive expected value. But even if they
did, there should be no correlation between the number of rejected lotteries and the
stopping decision in the experiment (as long as an individual's effort cost is uncor
related with
their risk aversion) because
the stopping decision is only determined
effort cost function. Status quo-based
loss aversion also predicts
by an individual's
28
If we pool the three treatments that are equivalent according tomodels of reference-dependent preferences
and R) and regress accumulated earnings or time spent working on a treatment dummy forHI (as
(LO, NOSAL,
inTable 1), thep-values of the treatment coefficient all drop below 0.01, the highest p-value being 0.003. Mann
Whitney (/-tests of these pooled treatment differences yield p-values below 0.001.
29The number of rejected lotteries increases in the degree of loss aversion regardless of the definition of the
reference point. This is clearly truewhen the status quo of zero is the reference point. It is also true for an expec?
tation-based reference point: Using Koszegi and Rabin (2007) choice acclimating personal equilibrium, as we did
to derive the predictions in Section II, there is a cutoff value for Y below which rejecting the lottery is the preferred
personal equilibrium. The cutoff increases in r\and A, i.e., a more
loss averse subject rejects more
lotteries.
This content downloaded from 131.220.108.137 on Mon, 18 May 2015 12:21:21 UTC
All use subject to JSTOR Terms and Conditions
APRIL 2011
THEAMERICANECONOMIC REVIEW
488
stopping behavior, nor do salience or reciprocity explanations.
subjects learned about the lotteries only after they stopped working;
Additionally,
thus the anticipation of the lotteries could not have influenced stopping behavior
no correlation with
directly.
If subjects have expectations-based,
reference-dependent preferences, however,
we should expect a positive correlation between the number of rejected lotteries and
the stopping decision: themore loss averse a subject is, the closer to/they should
stop. To see this, remember that the first-order condition for earning levels below
/is c'{e*)
=
w/2 + wrj(\
?
l)/4. If rjor A increase, i.e., if a subject getsmore
closer to /. For earning
loss averse, the optimal effort increases and thus moves
?
?
an increase in rj or A
the condition is c'{e*) ? w/2
levels above/
wry(A
l)/4;
decreases optimal effort,moving it closer to/.30Our data support this hypothesis:
RESULT
6: Subjects who are more
loss averse
stop significantly closer
tof.
regressions inwhich we regress the absolute distance
to
fixed payment on a subject's degree of loss aver?
the
earnings
sion and various controls. Data from all four treatments where reference-dependent
preferences could impact stopping behavior (LO, HI, NOSAL,
R) are included.31
We find in all specifications that the proxy for loss aversion is significantly nega?
Table
5 shows results of OLS
of accumulated
tive, i.e., more loss-averse subjects stop closer to the fixed payment.32 This provides
additional, direct support for reference-dependent preferences and loss aversion as a
in themain treatment difference.
key mechanism
V. Conclusion
In a simple real-effort laboratory experiment, we tested theories of reference
dependent preferences that assume the reference point to be a function of individual
subjects work
expectations. Our data is in line with predictions of these models:
more when expectations are high, and many subjects stop when piece rate earnings
equal the fixed payment. Three additional treatments ruled out alternative expla?
nations based
lottery choices
on salience
and reciprocity. We
that reference-dependent
also provided direct evidence from
preferences drive our results. Our results
30The model by Koszegi and Rabin (2007) also makes a more subtle prediction, which could generate this cor?
relation for reasons related to endogenous background risk. If subjects stop far from/they could try to "hedge" the
resulting risk by accepting more lotteries. Since the lotteries are independent from the draw of the envelope, this can
on special assumptions about choice bracketing, how?
(in some cases) reduce a potential loss. This prediction relies
ever, that are not very plausible. Namely, we would need to assume that subjects do not consider risk from outside
the experiment but that they do consider risk from themain stage of the experiment when deciding on the lotter?
ies. Numerous studies (e.g., Tversky and Kahneman
1981, Rabin and Georg Weizs?cker
2009) show that subjects
usually bracket more narrowly, and do not even bracket together two different lotteries that are shown on the same
sheet. Either way, only Koszegi and Rabin (2007) model predicts this correlation and other explanations
salience or non-reference dependent risk aversion) do not.
(e.g.,
31
Two out of 240 subjects chose lotteries inconsistently: they switched more than once between the safe and the
risky option which makes it difficult to interpret their choices. These two subjects are excluded from the analysis.
Results are unchanged ifwe include them.
32In SAL we expect there to be no relationship between loss aversion and stopping at three, given that stopping
at three does not help avoid losses. Indeed, ifwe estimate the same regressions as inTable 5 for SAL, without con?
decision
trols the point estimate is close to zero ( ?0.038
) and far from significant. Adding
zero, and is even positive in the final specification.
controls, the effect gets closer to
This content downloaded from 131.220.108.137 on Mon, 18 May 2015 12:21:21 UTC
All use subject to JSTOR Terms and Conditions
VOL.
101 NO.
2
ABELER
Table
5?Impact
ETAL.:
REFERENCE
of Loss Aversion
POINTS
on Distance
AND EFFORT
to Fixed
PROVISION
489
Payment
I
we-/|
Loss aversion
-0.489**
(0.220)
_(1)_(2)_(3)_(4)
-0.500**
(0.222)
Productivity
-0.518**
(0.222)
0.013
(0.009)
1 ifFemale
-0.472**
(0.236)
0.014
(0.010)
-0.188
(0.578)
Controls for treatments
No
Yes
Yes
Controls for temperature
No
No
No
Yes
Controls for time of day
No
No
No
Yes
Constant
Yes
6.040***
6.726***
7.522***
7.368***
(0.934)
(1.050)
(1.191)
(1.273)
Observations
238
238
238 238
Adjusted/?2
0.02
0.01
0.02
0.01
Notes: The table reports estimates fromOLS regressions. The dependent variable \we? f | is
the absolute distance of accumulated earnings to the fixed payment (in euro). Data from LO,
and R treatments are included in the analysis. Two subjects who chose inconsis?
HI, NOSAL,
tently in the lotterymeasure are excluded. The proxy for loss aversion is the number of lotter?
ies that a subject rejected (see text for details). The proxy for productivity is the time subjects
needed per table during the first stage (in seconds multiplied
parentheses.
***Significant at the 1 percent level.
**
Significant at the 5 percent level.
*
Significant at the 10 percent level.
by ?1).
Standard errors are in
thus contribute to understanding what determines
models
which
assume
the reference point. They support
the reference point to be formed by expectations, like Bell
(1985), Loomes and Sugden (1986),Gul (1991), orKoszegi andRabin (2006).
Our results are also relevant for the literature on reference points and labor supply.
Studies in this literature use field data on worker effort choices to testwhether the
response of effort to changes in incentives is consistent with the standard intertemporal
substitution of labor and leisure, or ratherwith loss aversion around a daily reference
income. In this literature the reference point has typically been treated as an unob?
served, latent variable. Camerer et. al. (1997) demonstrated that the daily labor supply
ofNYC cab drivers is in line with loss aversion around a daily income target.Camerer
et al. (1997) andF?rber (2005) bothnote,however,thatdaily earningsvarytoomuch
to be explained by a fixed daily income target. Partly in response to this evidence,
a theory of expectation-based,
and Rabin
reference
Koszegi
(2006) developed
a
to
the
that
allows
income
differ
in
target
dependent preferences
predictable way
across days. Our experiment adds to this literature by making the rational expecta?
tions about earnings known to the researcher, and by providing exogenous variation
while keeping other potential reference points constant. As noted by Koszegi
and
Rabin (2006), and subsequentlyshownbyCrawfordandMeng (forthcoming)
using
on expectations, antici?
(2005), if reference points are based
not
distort behavior relative to standard theory,
pated changes in incentives should
to
reflect
the anticipated change. For example, if an
that
given
expectations adjust
individual expects the hourly wage to be low on a given day, earning a small amount
the dataset of F?rber
This content downloaded from 131.220.108.137 on Mon, 18 May 2015 12:21:21 UTC
All use subject to JSTOR Terms and Conditions
APRIL 2011
THEAMERICANECONOMIC REVIEW
490
does not feel like a loss. But if the hourly wage is unexpectedly low, this does feel like
a loss relative to expectation, and can induce workers to work even harder to try to
reach their expectation, contrary to the standard prediction on intertemporal substitu?
implies thatworkers should decrease effortwhen thewage is temporarily
low. This distinction helps reconcile some of the seemingly conflicting findings in
the field evidence. Our results are complementary, providing controlled evidence that
tion which
expectations can in fact act as a reference point, and can affect effort provision.
An interesting direction for future research is to distinguish between different
expectation-based models of reference-dependent preferences. Our treatments are
not designed to test which way of specifying the reference point in expectations
is
the empirically most plausible: assuming that the reference point is themean of the
expectedoutcomes (like inBell 1985,Loomes and Sugden 1986, orGul 1991) or
distribution of expected outcomes
of these assumptions predict
earnings equal the fixed payment.
that the reference point is the whole
and Rabin 2006, 2007).
(like in, e.g., Koszegi
a higher probability to stop when accumulated
assuming
Both
experimental design provides a useful platform for pursuing this question in
if
the future, however, and could be extended to distinguish between these models:
a
over
are
two
distinct
fixed
determined
final
payments
by
lottery
subjects'
payoffs
and accumulated earnings, rather than just one fixed payment and accumulated earn?
Our
ings as in the current study, then predictions
are different across models.
Models
like theone ofLoomes and Sugden (1986) predicta higherprobabilityto stopwhen
earnings equal themean of the two fixed payments but not when they
one
two fixed payments. Models
the
like Koszegi
and Rabin (2006) predict
of
equal
a higher probability to stop at the two fixed payments but not at themean.
accumulated
REFERENCES
and Social Decisions."
Econometrica,
Akerlof, George A. 1997. "Social Distance
in Decision
under Uncertainty."
Bell, David E. 1985. "Disappointment
Making
33(1): 1-27.
Bernheim,
Research,
Journal of Political Economy,
841-77.
1994. "A Theory of Conformity."
102(5):
and Lorenz Goette.
2009. "Performance
Jeffrey Carpenter,
Pay and Worker Coopera?
from an Artefactual
Field Experiment."
Journal of Economic
Behavior
and Organi?
B. Douglas.
Burks, Stephen,
tion: Evidence
zation, 70(3): 458-69.
Camerer,
1005-27.
65(5):
Operations
Colin,
Linda
Babcock,
George
and Richard
Loewenstein,
Thaler.
1997.
"Labor
Supply
of
New York City Cabdrivers:One Day at a Time." QuarterlyJournal ofEconomics, 112(2): 407-41.
and the Construction
1999. "Anchoring, Activation,
of
B., and Eric J. Johnson.
and Human
Behavior
Decision
115-53.
Processes,
Organizational
79(2):
and Doron
An Inter?
Sonsino. 2007. "Social Distance
and Reciprocity:
Charness,
Gary, Ernan Haruvy,
net Experiment."
Journal of Economic
Behavior
and Organization,
88-103.
63(1):
Chapman,
Values."
Gretchen
2007. "Consistency
and Hetero?
Fisman, Douglas
Gale, and Shachar Kariv.
Syngjoo, Raymond
American
under Uncertainty."
Economic
1921-38.
Review,
geneity of Individual Behavior
97(5):
of Labour
from Taxi Drivers
in
Chou, Yuan K. 2002.
"Testing Alternative Models
Supply: Evidence
Economic
17-47.
Review, 47(1):
Singapore
Singapore."
Choi,
and Juanjuan Meng.
"New York City
Crawford,
Vincent,
Forthcoming.
Revisited:
Preferences
with Rational-Expectations
Reference-Dependence
Economic
Review.
Income." American
Doran,
Kirk
Bennett.
2009.
"The Existence
and Position
tionsforDaily Labor Supply."Unpublished.
Falk, Armin,
and Urs Fischbacher.
54(2): 293-315.
2006.
"A Theory
of Daily
of Reciprocity."
Income
Games
Cabdrivers'
Targets
Reference
Labor
Supply
and
for Hours
Points:
and Economic
This content downloaded from 131.220.108.137 on Mon, 18 May 2015 12:21:21 UTC
All use subject to JSTOR Terms and Conditions
Implica?
Behavior,
VOL.
101 NO.
2
ABELER
and Andrea
Falk, Armin,
Ichino.
ET AL.: REFERENCE
"Clean
2006.
POINTSAND
EFFORT
on Peer Effects."
Evidence
F?rber, Henry S. 2005. "Is Tomorrow Another Day? The Labor Supply of New
Journal of Political
Economy,
113(1): 46-82.
and Labor Supply:
Preferences
F?rber, Henry S. 2008. "Reference-Dependent
Economic
1069-82.
Review,
City Taxi Drivers." American
98(3):
Field
Randomized
American
Experiment."
Urs. 2007.
"Z-Tree:
Zurich
Fischbacher,
mental Economics,
171-78.
10(2):
Economic
Toolbox
Review,
for Ready-Made
Econom?
of Labor
Journal
ics,24(1): 39-57.
"Fairness
and Retaliation:
The
2000.
Fehr, Ernst, and Simon G?chter.
Journal of Economic
159-81.
Perspectives,
14(3):
2007. "Do Workers Work More
IfWages
Fehr, Ernst, and Lorenz Goette.
491
PROVISION
York City Cabdrivers."
of New
The Case
Economics
of Reciprocity."
Are High?
298-317.
97(1):
Economic
York
from a
Evidence
Experi?
Experiments."
for Experimenters:
Eco?
1998. "Good News
Frank, Bj?rn.
Subjects Do Not Care about Your Welfare."
nomics Letters, 61(2):
171-74.
and Eric J. Johnson. 2007. "Individual-Level
in
Loss Aversion
G?chter,
Simon, Andreas
Herrmann,
Riskless
and Risky Choices."
Economics
Discussion
mental
of Nottingham
University
Paper 2007-02.
Centre
for Decision
Research
and Experi?
and Christopher
Genesove,
David,
the Housing Market."
Quarterly
Ben.
Greiner,
2001. "Loss Aversion
from
and Seller Behavior:
Evidence
Mayer.
Journal of Economics,
1233-60.
116(4):
In Forschung
"An Online Recruitment
und
System for Economic
Experiments."
f?r wissenschaftliche
Rechnen. Gesellschaft
Bericht, Vol. 63,
Datenverarbeitung
2004.
Wissenschaftliches
ed. Kurt Kremer
and Volker Macho,
79-93.
G?ttingen,
Germany:
f?r wissenschaftli?
Gesellschaft
che Datenverarbeitung.
1991. "A Theory
Gul, Faruk.
of Disappointment
Aversion."
667-86.
Econometrica,
59(3):
and Frauke Lammers.
2008. "The Role of Expectations
in the Formation
of Reference
Hack, Andreas,
Points over Time." Unpublished.
Heath,
Chip, Richard
P. Larrick,
and George
chology,38(1): 79-109.
Heidhues,
Paul,
Loss Averse."
and Botond
American
Koszegi.
Economic
2008.
Review,
Wu.
1999.
Karen
E.,
and Daniel
Kahneman.
Daniel,
and Amos
Tversky.
1995.
1979.
Cognitive
Psy?
when
Consumers
Are
"Measures
2010.
Schemes:
Moral
"Binary Payment
2451-77.
100(5):
in Estimation
of Anchoring
Tasks." Per?
sonalityand Social PsychologyBulletin, 21(11): 1161-66.
Kahneman,
Points."
and Price Variation
"Competition
1245-68.
98(4):
Daniel M?ller,
and Philipp Weinschenk.
Fabian,
Herweg,
Hazard
American
and Loss Aversion."
Economic
Review,
Jacowitz,
as Reference
"Goals
"Prospect
263-91.
47(2):
and Richard
Jack L. Knetsch,
Kahneman,
Daniel,
ment Effect and the Coase
Theorem."
Journal
Theory:
An Analysis
of Decision
under Risk."
Tests
of the Endow?
Econometrica,
H. Thaler.
of Political
"Commitment
2009.
K., and Julia Nafziger.
Koch, Alexander
of Labor Discussion
Paper 4049.
and Matthew
Rabin.
2006. "A Model
Botond,
Koszegi,
terlyJournalofEconomics, 121(4): 1133-65.
Koszegi,
Botond,
and Matthew
Rabin.
nomicReview, 97(4): 1047-73.
and Matthew
Rabin.
Botond,
Koszegi,
Economic
909-36.
Review, 99(3):
Locke, Edwin A., and Gary P. Latham.
wood Cliffs, NJ: Prentice-Hall.
Locke, Edwin A., and Gary P. Latham.
and Task Motivation:
A
1990.
1325^4-8.
Institute for the Study
of Reference-Dependent
Preferences."
2007.
"Reference-Dependent
Risk
2009.
"Reference-Dependent
Consumption
1990. A Theory
of Goal
Setting
Attitudes."
American
Plans."
Theory
705-17.
1987. "Testing for Regret and Disappointment
Sugden.
118-29.
Economic
Journal, 97(S):
Uncertainty."
to Realize
Their Losses?"
1998. "Are Investors Reluctant
Journal
Odean, Terrance.
Setting
in Choice
in Choice
of Finance,
1775-98.
Rabin,
Matthew.
Rabin,
Matthew.
1993.
"Incorporating
nomicReview, 83(5): 1281-1302.
2000.
"Risk Aversion
metrica, 68(5): 1281-92.
Fairness
into Game
and Expected-Utility
Theory
and Economics."
Theory:
A Calibration
American
Theorem."
This content downloaded from 131.220.108.137 on Mon, 18 May 2015 12:21:21 UTC
All use subject to JSTOR Terms and Conditions
Eco?
Engle
of Goal
Consistency
Quar?
American
and Task Performance.
a Practically
Useful
"Building
American
Odyssey."
Psychologist,
57(9):
1986. "Disappointment
and Dynamic
Sugden.
271-82.
Studies, 53(2):
of Economic
2002.
35-Year
and Robert
Loomes,
Graham,
under Uncertainty."
Review
and Robert
Loomes, Graham,
"Experimental
Economy,
98(6):
to Self-Rewards."
under
53(5):
Eco?
Econo?
492 THEAMERICANECONOMIC REVIEW
and Georg Weizs?cker.
2009. "Narrow Bracketing
1508-43.
Review,
99(4):
International
2000. "Loss Aversion
Shalev, Jonathan.
Equilibrium."
269-87.
Rabin, Matthew,
can Economic
Tversky,
Amos,
Tversky,
Amos,
Whynes,
David
and Daniel
Kahneman.
Science, 185(4157): 1124-31.
and Daniel
Kahneman.
Choice." Science, 211(4481): 453-58.
K., Zoe
Philips,
and Emma
Economics, 14(11): 1191-95.
1974.
"Judgment
1981.
Frew.
"The
2005.
APRIL
and Dominated
Journal
under Uncertainty:
Framing
"Think
of Decisions
of a Number...
2011
Choices."
of Game
Heuristics
and
Theory,
Ameri?
29(2):
and Biases."
the Psychology
Any Number?"
This content downloaded from 131.220.108.137 on Mon, 18 May 2015 12:21:21 UTC
All use subject to JSTOR Terms and Conditions
Health
of