Time-inconsistent Preferences

and Consumer Self-Control

STEPHEN J. HOCH
GEORGE F. LOEWENSTEIN”

Why do consumers sometimes act against thair own better judgment. engaging
in behavior thai is often regretted after the fact and that would have beBn
rejected with adequate Jorethought? More generally, how do consumers
attempt to maintain selI•controI in the (ace of time-inconsistent preferences? D
This article addresses con- sumer impatience by developing a decision-theoretic o
model based on reference poims. The model explains how and why consumers w
expeñence sudden increases in desire for a product. increases that can result in nl
the temporary overriding of long- term preferences. Tactics lhat consumers oa
use to control their own behavior are also discussed, Consumer seifmontrol de
iS framed as a struggle between two psy- chological lorces. desire and d
willpower. Finally, two general classes of self-controi strategies are descnbed fro
those that direL;tly reduce desire. and those that overcome desire through m
willpower. htt
ps
://
ac
ad
The same person will vary in his mood, e
being at one t ime impatient, and greedy model. Consumers decide whether to make a mi
for present enjo ymc nl; while at a not he r purchase, and which purchase to make, by weighing c.
his mind dwells on ltte futu re. and hc is the costs and benefits of alternatives. The terminology ou
used to describe these decisions—“trade-offs,” p.
willing to posipone all enjoyments that can
co
convc niently be made to wait. Sometimes “attributes,” “choices,” “decision rules”—reveals a m/
he is in a mood to care little for anything particular view of behd vjor. Consumers are seen as jcr
else: soiiieti mes he is like the chil d ren dispassionate in- formation processors. evaluating /ar
who pick the pt u ms o ut of their puddi ng ticl
to cat them at once, somei i mcs like ihose
alternatives in a boundedly rational fashion and
e/
who put them aside to be eaten last. effortlessly imple- menti rig decisions (Bett man 17
[Marshall 1 890, p. 100) 1979). I ndeed. m uch purchase behavior is probably /4/
Who among us has never been gu iJiy of well described by these utilitarian terms. But, as 49
being u nable to deny himself lhe momen- suggested in the earlier ob- ser nations of Marshall and 2/
tary enjoy mc ni of thut 1 uscious dcsscrt or Bfi h m-Bawerk, a more complete understanding of 17
the fragrant Havana thai the doctor or- 97
consumer hehavior m ust recognize that people are
dered us to La rgo? And that in spite of 24
influenced both by long-term rational concerns and by 3
kn owing pe rfcrtl y we l1 that it is going to
more short-ierm emotional factors. by
bring an aggravation of our ail ment which
su bseq ucn1 unbiased appraisal will rec- Holbrook, O’Shaugh nessy, and Bell ( 1990) write, gu
“Recent approaches to consumer research have tended es
ognize as being far more u n pleasant than
to regard corisu mer behavior as a mode of reasoned t
the renu nciatio n of that iri fling enjoy- on
ment. [Biihm-Bawerk ( 1898) 1959, p. 269] action or as a repository of emotional reactions . 09

I n marketing, the dominant paradigm used to de-

scribe consumer behavior is ltte rational choice
[but] a one-Sided focus on either aspect by itself—ac-
tions or reactions—provides a distorted view of the
consumption experience.” We agree. There is a clear
O

Stephen J. Hcch professor of marketing and behavioral science

and Georg r. Loewc nstein is associate profcsso r of hcha vio ral sci-
ence, both ai the Center for Dccisio n R cscarch. Graduate School
of Busi ness, U niversii y of Chicago, i 10 i E. 5 8th St.. Chicago, I L
60637. Both authors contributed equally to the w rising of this paper.
This work was supported in part by the Beatrice Companies, I nc.,
Faculty Research Fu nd at the Graduate School of Business, U ni-
versity of Chicago, a National Science Fo undalion grant #SES-
891075 5 to S.J.H ., and bt' grants from they Russell Sage Foundation
and the Alfred P. Sloan Foundation to G.F.L.
492 modeled formally
need to examine the interaction between rational and
hedonic motives—between what Albert H irschman
( 1977) called the “interests” and the “passions” and
what Abelson ( 1963) labeled “cold” and “hot” cog-
nition. Although each perspective adequately
describes a wide range of consumer behaviors, neither
alone can provide a complete account. One reason that
little progress has been made in integrating the
rational with the hedonic is an incompatibility in
vocabulary. Typ- icall y, the rational side has been

<c 1991 oy JOt4 RNn I. OF CONSUMER RESEARCH, Inc. e Vol. l7 e March l99t
All n$tMs l<served. fXt9T3 30 I /9 I /1704-OXI4$tI2.X

CONSUMER SELF-CONTROL

while the hedonic side has bcen described more qual-
itatively. One way to lie the cognitive to the emotional
is to place them in a framework with a com mon lan-
guage.
We do not intend to provide a com plete model of
the interplay betwecn cognition and emotion. Our in-
terests here are narrower and our efforts m ore modest.
We oifer an eco nomic-psyc hological model that seeks
io integrate ihe rational and emotional forces influ-
encing consumer self-control. Although consumers are
often even handed in their balancing of present and
future benefits and costs, certain situations can induce
extreme impatience. At such times, buying assu mcs a
more hedonic character (H irsch man and Holbrook
1982), one better characterized by sudden increascs in
desire accompanied by inner conflict and attem pts to
control one’s own behavior. Shi1’ts in time perspective
of this type have becn labeled “myopic” or “time-in-
consistent” by economists (Strotz 19561 and “impul-
sive” by psychologists (A inslie 1975). Self-control, as
we defi ne it, consists of efforts on lhe part of’ the con-
sumer to avoid or resist behaving in such an inconsis-
tent man ner.
TIME-I NCONSISTENCY
A time-inconsistent choice is one thal would not
have been made if it had been contemplaled from a
removed, dispassionate perspcctivc: it represents a
transient alteration in lastes, not a permanent reeval-
nation of an alternative due to receipt nf new infor-
mation (cf. Stigler and Becker 1977). Time-inconsis-
tent behavior is of special interest to economists
because of its significance for macroeconomic policy
(Kydland and Prescott 1977) and sari rigs behavior
(Thaler and Shefrin l9fi 1). Psychologists have been
concerned mainly with the connection between im-
pulsi vity and dieting, addiclion, and other self-regu-
lation problems. The ability to delay gratihcation has
been studied extensively in developmental and clinical
settings (M ischel 1974).
Time-inconsistcncy raises inn po rtant q uestions
about consumcr sovereignty (Slrotz 1956). Are im-
patience-driven decisions as legitimate (normative) as
the even handed preferences that are usually assu med
to reign? Is there no arguing with t astcs, as
economists often assert? The consensus of social
scientists, and apparently of consumers them set yes,
is that time-in- consistent preferences are not as
legitimate as their more farsighted counterparts.
Consumers ialk about being driven to buy something
against their own better judgment, and, as a result, we
com monly observe at- tempts by consumers to
control their own behavior, to avoid or resist episodes
of extreme impatience. Popular magazines aid
consumers in control efforts by offering elaborate
instructions on how to fight the “urge to splurge” and
avoid “binge buying” ( fimr 1987).
This article begins with a discussion ef consumer
time preference and the causes of time-inconsistency.
494
it i3 a m omentary and involuntary departure from the
consumer’s dominant preferences. Even while behav-
493
We propose a m odel to explain consumer impatience
and the circumstances under which it will occur. Spe-
cifically. the model addresses the question of how and
why consumers experience “sudden, often powerful
urge(s) to buy something immediately” (Rook 1987)
that can in vol ve temporary overriding of long-term
preferences. Next, we discuss tactics that consumers
use to control I heir own behavior and to avoid time-
inconsistency. We organize our discussion around a two-
factor modcl or behavior, where the battle for self-
control is conceptualized as a struggle between the two
psychological forces o1’ desire and willpower. We dis-
tinguish t wo classes of self-control strategies: (a) those D
ihal attempt to directly reduce desire, and (h) those o
w
lhat seek to ovcrcome desire through a variety of will-
nl
power tactics.
oa
Before continuing, it is important to place this work de
in proper perspective. First, we assume that consumers d
do i ndeed have long-term preferences. We do not ad- fro
m
vocate (or require) the extreme position of some econ-
htt
omists, who argue that tastes are stable and essentially
ps
constant across all consumers, Rather we assume that ://
consu m ers have I eng-terms goals that, in turn, guide ac
their preferences (M iller, Galanter. and Pribram 1960). ad
Second, we assu me that passionate behavior can be e
mi
modeled. We retain the essential cost-benefit orien- c.
tation of the utilitarian perspective, bui m odify it to ou
incorporate the emotional reactions excluded from p.
most economic models. Finally, our approach is most co
releva nt to contemporary Western consu mers, where m/
jcr
possessions play an indirect role in defining individ- /ar
ually bascd identities, and less relevant to more tra- ticl
ditional societies, where group identity is the guiding e/
force behind self-concept (Belk 1984). 17
/4/
49
A RE FERENCE-POINT MODEL 2/
OF DESI RE 17
97
W hy do people sometimes act incOn5istently over 24
time engaging in behavior that would have been re- 3
by
jecte‹l if contem plated in adva nce and that may be gu
regretted after the fact? This question has puzzled gen- es
erations of social scicntists. One view, espoused by t
Marshall ( 1 890) in the openi rig quote, is thai time- on
inconsistent behavior arises from unpredictable 09
O
changes in moods and tastes (also see Bass 1974). I n
a modern rendition of this view, Winston ( 19801 pro-
posed an econ omic model of tmpulsivity in which a
decision maker randomly vacillates between two sets
of preferences, one myopic and the other farsighted.
Winston descrilres a variety of strategies that people
can employ to maximize expected utility in the face
of inconsistent preference fluctuations. This economic
model of impulsivity has two shortcomings. First, it
placcs time-inconsistent preferences on an equal foot-
ing with temporally consistent preferences. But im-
patience is not simply the random manifestation of an
alternate and equally valid preference ordering; rather,
ing impulsively, consumers often view their own de-
cision making as distorted (Rook 1987). Second, the
random-preference model fails to shed light on the ba-
sic qucstion of when and why m yopic preferences
come into play. JOURNAL OF CONSUMER RESEARCH
An alternative and better-k nown cx planation of im-
patience (Ainslie 1975; Strotz 1956) is based on lhe FIGURE 1
economic concept of discounting. According to t his
view, people are disproportionately attracted to im- ADAPTATION ANO THE VALUE OF CONSUI\1PTION
mediatel y available rewards. When two rewards (e.g. , Sati slactio n
the pleasure of a cigarctte and thc plcasu re ot good
health) are both substantially delayed, the individual
is able to make a rational trade-off between them ,
However, when one reward (the cigarette, in this ex-
ample) is imminent, it exerts a disproportionate at-
traction. The discounting perspective is accurate as
far as it goes, but it leaves basic questions unanswered.
Why are certain types of goods (e.g.. dessert) associated
with impulsivity and not others (e.g., gasoline)? Dis-
counting also fails to explain why other factors, such D
o
as physical proximity. are com mt›nly associated with w
time-inconsistency. nl
Although recognizing the i mporlance of ti me (as oa
well as other dimensions of pro ximit y), our model of de
impatience differs from the eco nomic discounting d
fro
perspective. We believe that time-inconsistent pref- m
erences arc due to suddcn i ncreascs in desire brought htt
o n by a shift in the consumer's reference point. Ref- ps
erence-point shifts can be precipiialed by any number ://
of factors (e.g.. physical proximity in a store), causing ac
ad
the consumer to partially adapt to I he notion of owning e
or consuming the product. After a reference point mi
shifts, consumers not only attach positi ve utility to t he c.
object itseir(what o ne might call a standing or existing ou
desire before the reference-point shi ft). but they also p.
Tversky’s ( 1979) prospect theory and more recently co
attach negative utility to fai lurc to consu me the object. m/
Failure to purchase implies more than the forgone applied lo consu mer choice (Thaler 19SO, 1985) and jcr
pleasure fro m consu m ption; the individual actuall v intertemporal choice (Loewenstein 1988). The value /ar
feels deprived, This deprivation is a psychological state function represents the consumer’s satisfaction level ticl
of need analogous to physi ological states of need exclusive of any costs (the j -axis) as a function of a e/
(which can range from h unger and thi rst to addiction). discrete outcome levcl (purchasc or nonpurchase) rel- 17
ative io some refcrencc point. The reference poini is /4/
49
Reference Points and Deprlvation located wherc the value function intersects the .v-axis. 2/
The concept of a reference point, or adaptation level, T he val ue fu nction is assu mcd to be steeper for neg- 17
has a long history in psychology. Reference points play ative departures (losses) than for positive departures 97
a central role in research on perception (Helson 1964), (gainsl from the rcferc ncc point (Bell 1982, 1985; 24
Fish burn 197 7; Kah neman and Tversk y 1979). 3
satisfaction with achicvcment (Lewin et al. 1944), dis- by
tributive justice (Walster, Bcrscheid, and Walster Consider a good that would yield a vaI uc of N, if it gu
19 73), saving behavior (Duesenberr y 1949), and lifc were purchased. The no-adaptation value function, es
satisfaction (Brick man, Coaies, and Janoff-Bulman rcpresented by the lighter line lfiat goes through the t
1978). The reference point reflects the fact that people origi n, is d rawn such that the purchase of the good on
yields D , and nonpurchasc yields zero. This no-ad- 09
are less concerned with absolute attainments than with O
aptation function depicts thc case of a consumer who
attainments relative io some psychologically relevant
would dcrive salisfaction from owning an object (Erased
comparison point.
on long-term needs) bu1 who has not adapted to the
Figure 1 illustrates the effect of a refcrence-point
notion of possessing it. Wiih no adaptation, the stecpl y
shift using a framcwork introduced in Kah neman and
sloped negative region of the value function is not rel-
evant.
The value function’s usefulness lies in showing what
happens whgn the consumer partially adapts to pos-
session of 1 hc good. The partial adaptation casc is de-
picted by the bold-line value function. As com pared
with the no-adaptatio n situation. thc zero point of the
value function has movcd to the right, signifying that
the consumer is now in an intermediate state between
CONSUMER SELF-CONTROL stead results in deprivatio n (the distance bet wccn the
origin and where the bold value function intersects the
owning and not owning the prod uci. Failing to make -axis). Now, overall desire for the product is repre-
the purchase is no longer affectively neutral, but in- sented by the distance +2 ( e., the utility resulting from
consumption coupled wilh the rclief ensuing from not
feeling deprived).
In both the no-adaptation and adaptation cases. we
define desire D and D2l as the difference between
owning and nor ohm ng the object. The graphic de-
pictio n of desire D) in Figure 1 can be written in
equation form.
D — o(P — r) + g(r — 0), (1)
where a is the slope of the value function in the positive
region, fl is its slope in lhe negaii ve region, and r is the
consumer’s level of adaplalion between purchase (N)
and nonpurchase (0). The value N is normalized so
that a P iS equal to the consumer’s no-adaptation level
of desire for the good D .
As long as fi > n, an increase in r will raise D.
Desire represc nts the consu mc r*s levcl of
motivation to pos- sess the object; in the no-
adaptation case, it is based solely on a standing
preference for ihe iiem. Because the consumer views
the purchase as hypothetical, fail- ure to purchase
does not result in deprivation (the consu mer
possesses nothing and, so. loses nothing). The consu
mer will buy lhe product whenever desire exceeds the
costs (both economic and psychic) of pur- chase.
Deprivation, induced by partial adaptation. in-
tensifies desire and the motivation to purchase
because
> Nt . The consumer who partially adapts ha5 al-
ready had a taste and may drcad the pain of having to
forgo further tastes. Similarly, the bait-and-switch vic-
tim, who anticipates making a purchase but fi nds the
item out of stock, will feel deprived and may feel com-
pelled to purchase a substitute item to mitigate the
immediate frustratio n.
Divergent social science research points to the
pow- erful incentive value of negative departures
from a ref- erence level. Research on a nima I
learning has dist in- guished three types of incentives:
reward, punishment, and frustrative non reward.
Frustrati ve nonreward oc- curs “when we fail to
reinforce a response that has previously been
reinforced” (Skinner 1950, p. 203). There is evidencc
that frustrativ e n on reward has greater incentive
value than either reward or punish- ment (Amsel
1958; Wagner 19 5 9). For example, the powerful
motivating effect of intermittent reinforce- ment can
be explained in terms of the frustration brought on by
nonreinforced trials (Amsel 1962). Re- search on
social un rest also provides evidence for the
significance of negati vc departures from reference
lev- els. Most theories assume that discontent results
from a negative discrepancy between current
position and some reference value, whethcr the
reference value is
495
495
expectations about the future or the consumption level
of a referent individual or group (Austin 1977; Gurr
1970; W alster et at. 1973). Falling below onc’s refcr-
ence point or adaptation level, whether biological or
psychological, can lead to feelings of deprivation and
provides lhe moi ination to improve one’s relative po-
sition.
E ynam it E’ff‹•ei.‹ vf R eference- Pnint Shi fts. Al-
though the rcferencc-point shift depicted in Figure 1
provides a mec han ism for the motive force behind
time-inconsistency, it is a stat ie conception. Most de-
cisions madc in the present have hedonic consequences D
o
that extend over time. If desire is not t ransient, the
w
loss depicted on the value fu nction will understate the
nl
deprivation anticipai cd or cxperienced by the con- oa
su mer while consumption is delayed. The deprivation de
d
induced by a reference point shift is not instantaneous,
fro
but may linger. Thus. the relevant loss must take into
m
account thc lc ngt h of time during which the consumer htt
feels deprived. Figure 2s depicts the choice between ps
an in ferior, inn mediate object and a superior, delayed ://
ac
object under conditions of partial adaptation. The
ad
smaller rectangle in Figure 2 represents the utility de- e
rived from consuming a smaller immediate reward; mi
the larger rectangle depicts lhe utility derived from a c.
larger delayed reward. The area marked “deprivation ou
p.
from waiting” represents the negative utiJ ity experi- co
enced while waiting for the larger delayed reward. The m/
delayed reward could be a concrete object (e.g., a car) jcr
or something more ahstract (e.g., long-term health /ar
ticl
benefits). As Figure 2s shows, the more deprived one
e/
feels while waiting, the greater is the incentive to con- 17
sume quickly so as to terminate the stream of negative /4/
utility. 49
In certain cases, thc consumer may be aware that 2/
17
desk re for an object is transient and may opt for the 97
delayed option, k now ing that the deprivation will be 24
short-lived. This situation is depicted in Figure 2b, 3
with time, depri vation from waiting decreases. Often by
de- privation decreases with time because the gu
es
consumer moves on to other activities. Reduced t
attention to thc object of desire can result in a leftward on
shift of the reference point back toward the non 09
purchase zero point in Figure I . An example of O
decreasing depriva- tion over lime can arise in eating
behavior. Some peo- ple have learned ihat, even if
they remain ravenous after a big meal, they should
delay the decision to order dessert. The reason is that
changes in osmotic pressure in the stomach and blood
sugar that accompany eating, and reduce feelings of
hunger, do not occur immedi- ately (Guyton 1971).
Within 20 minutes, however, the pain of abstinence
usually fades. In other cases, desire may intensify
over t i me, a situation illustrated in Fig- ure 2r,
Again, physical appetites provide a context for
under5tanding such phenomena: feelings of hunger
and thirst lypically intensify if left unsatisfied. This
CHOICE BETWEEN IMMEDIATE AND DELAYED CONSUMPTION
ta) C a n s tant d e p rl v ali o n

(b) T ran s i ent d e p riva t i a n
may he one reason why ph ysical appetites are such
potent instigators of impulsivity: they just do not go
away.
The pain associated with not buying is a recurring
theme in consumers’ accounts of their own inn pulsiv-
CONSUMER BELF-CONTROL
subjects were less willing to delay. It appears that plac-
JOURNAL OF CONSUMER RESEARCH
ity. For example, consumers asked to discuss recent
impulse purchases (Rook and Hocli 1985) repeatedly
commented on the difficulty of deferring consum Ation.
“It gnaws at me until I buy it.” “The feeling staris
when 1 see something . . it comes on very quickly
and is a persistent nagging.” Thc deprivation experi-
enced by the consumer is, thus, analogous to a mild
addiclion. Addiction is a com plex phenomenon with
m ultiple stages, progressions, and in fi uences (Marlatt
et at. 198d). Bul some research suggests that the ad-
dict’s powerful urge to consume is not so much a search
fo r plcasure (addictive substances lose hedonic potency
with prolonged usc) as a flig ht from pain, the pain of
not consu ming (Sil vcrstein 1976; Solomon and
Corbit 1974).
I n su m, we believe that lime-inconsistent prefer-
ences are i nduced by reference-point shifts in which
consumers adapt to possession of not-yet-purchased
items. When reference points shi fl, the consumer’s de-
sire for the nonpurchascd object increases. More im-
portantly, the consu mer desires to purchase or con-
surnc the object as quickly as possible so as to terminate
the stream of deprivation. I n the next section, we dis-
cuss en viron mental faci ors that can lead to
reference- point shi fts and prom ole impulsivity.
Proximity-induced Impatience
l n gencral, refcrcnce-point shifts arc induced by a
process of accom m odation or adaptation. People ex-
posed to a persistcnt sensory stim ulus (c.g.. a noise or
smell) adapt to lhat slim ulus to the point at which they
are no longer aware of il: they are sensitive only to d
cviations from the adapted-to-stimul us level. Re-
garding co nsu mcr behavior, reference-point shifts are
caused by the consumer’s adapting to possession of a
good that has not been purchased. There are many
mechanisms that can cause such adaptation (a right-
ward shift of t he reference point in Fig. 1 ), All of the
mechanisms involve increasing lhe consu mcr's prox-
imity to thc good along somc dimension.
Ph vsi‹'al Nro. Uni/r. Perhaps the most potent in-
duccr of reference-point shifts is ph ysical proximity.
The effect cif physical proximity on impulsivity has
becn well documented in experiments on delay of
grat ifi cation by M ischcl ( 1974). In a typical experi-
ment, children are placed in a room and taught that
they can quickly summon the experimenter by ringing
a beIL They arc then givcn a choice between an im-
mediate inferior reward (e.g.. a single marshmallow)
or a delayed superior reward (two marsh mallows). I f
ihe child is able to wait for the experimenter to return
without ringing the bell, s/he obtains thc superior
item. I mpulsivity is measured by how long the subject
writs before ringing the be1L In an early experiment,
Mischcl and Gruscc ( 1967) examined the cffeci of the
visiblc presence of thc rcward on willingness to wait;
when eit her the immediate or delayed reward was in
view,
ing the object in view leads to a rcferc rice-point shift that
increases deprivation and makes waiting more dif- ficult.
A variety of marketing practices ma y induce a ref-
erence-point shift by increasing physical and sensory
pToximity. Much advertising shows the product in use,
increasing proximity by vividly simulating the cxpe-
rience (Macl nnis and Pricc 1987. Wclls 1987). Distilled
spirits ads opi for simple “boitle, glass. and ice” de-
pictions to help consumers visuahze the ex perience.
Marketers also use atmospherics t K otlcr 197 3— 1974;
Nord and Peter 1980) to provide the senses wilh a hint
of the real experience. For example, Mrs. Field’s
Cookies has successfull y overcome high rcnts by
pu mping chocolate-chip cookie odor into the heavy If-
trafficked corridors o1 A mcrican shopping malls.
Sometimes marketers actually gi ve thc consumer a
taste through sampling. Scratch and -sniff perfume
samples are accompanied by product order forms.
Publishers who send books and magazines into the
home with the cxplicit assurance that they can be re-
turned if they are not wan led may also shift Jhe con-
sumer’s reference point (Thaler 1980), as may’ sellers
of Oriental rugs who encouragc customers to borrow
rugs to “see how they look in your home.” Clearly
there are otten unanticipated transaction costs asso-
ciated with product return. howcvcr. it is also the case
that parting with a book that has graced one’s coffee
tablc. or a rug that has cnhanccd thc atmosphere of
one’s living room. is more dt the ul t than turning away
from the same book or rug displayed with cou ntlcss
others in ihc rctail outlct (K nctsch. Thaler, and
Kah neman 1987).
Temporal Pro.’irnii\!. Earlier theoretical accounts
of impulsivity (A inslie 1975: Strot z 1956) focused
on only one dimension of proximity, proximity in
time.
I n the animal-learning literature, it is well known that
the more immediate a reward, the greater is its rein-
forcing val ue (Chung and Herrnsicin 1967). Hence.
inn mediate availability of a reward will tend to in-
crease the desire for ii. Research on lime preference
suggests that increasing temporal proximity not o nl y
increases desirability. but also increases impatience.
Willingness to delay gratification in exchange for
greater rewards decreases as consumption ot›jects be-
come irn mingnt (A inslie and Hacndel 19 h 3; Benzion,
Rapoport, and Yagil 1987). The increase in impatience’
duc in part to nonexponent ial discounting (A inslie
1975), hut it is also exacerbated by a temporally in-
duced reference-point 5hift (Loewenstein 1988).
I n four experimcnts with five- to seven-year-old
children, Irwin, Armitt, and Simon ( 1943) examined
preference for im mediate- and delayed-choice objects.
Children were shown two toys. one to bc received inn-
mediatcly and one later; preferences for the two toys
were measured. The children overwhelmingly prc-
498
have been previous income levels (i.e., an example of
intraindividual social comparison and its ePect on im-
pulsivity).
497
fcrrcd the more lemporally proximate toy (75 percent),
whether faced with a one-wcek (b2 percent) or lh ree-
min ute (68 percent) delay in receiving the other toy.
L oewcnstein ( 1990) examined the effects or temporal
proxi m ity by manipulating the interval that high
sch ool sophom ores anticipated waiting before receiv-
ing a reward objecL Subjects either expected an early
(e.g., four weeks) or a late delivery (e.g.. eight weeks)
of a $7 gift certificate from a local record store. After
a two-week delay, subjects were given a choice between
$7 at the early date or $8.50 if they delayed until the
fare date. Subjects expecting an early delivery were less
I ikely to wait than lhose expecting a later delivery (42 D
percent vs. b4 pcrcenl). Apparently. subjects adapted to o
t hei r anticipated delivery dates, and those subjects who w
felt rem porally closer to the reward were more nl
oa
impatient than those who felt tcmporally farther away. de
Some marke ling efforts may increase impatience by scl d
ling inn mine nt opportunit y, reminding the con- su fro
mer that the prod uct is “yours for the asking” or “only a m
phonc call away.” Pizza-delivery services that give a so htt
ps
bsiantial discou nt if the pizza is not delivered within 3t) ://
minutes ntit only are guaranteeing a warm. nonsoggy ac
product. but are also playing on the time element in a ad
domain where impatience tends to be particularly e
severe. The di 5count raises the credibility of the claim mi
c.
and offsets any anticipated impatience once the pizza is ou
ordered. since late delivery confers Signifi cant p.
economiC Savings. Many direct-response ad co
vcrtisemcnls are accompanied hy a toll-free hot line m/
allowing inn mediatc ordering. Direct-mail catal og jcr
/ar
companies not only feature toll-free n umbers with 24- ticl
hour order-taking operators, they also provide ihe e/
pption of next-day express deli ver y for an extra fee. 17
Television-bascd marketers of carpeting offer next-day, /4/
in-home viewing of samples and 2—3-day dcli vcry 49
2/
thereaftcr. 17
So ’ill C.’oinporison. A third factor that appears 97
c apable of inducing a refcrence-point shift is social 24
com parison. People faced with an adverse social com- the 3
by
parison want what their more tortunaie peers aJ ready gu
possess, and they do not want to wait. For example, es
sociological rcsearch (Easlerlin 1974: Merton 1968; t
Stouffer et al. 1949) suggcsts that pcople or" all income on
levels tend lo compare themsel ves with others who are 09
O
slightly betlcr off lhan they are. Festingcr ( 1954) argucd
lhut people prefer to compare themselves to “superior”
others, for both informational and status reasons (cf. is
Brickman and Bul man 1977). A reference-point shift
induced by social comparison increases proximity and
may reduce willingneS5 to delay consumption. Related
research on coaCtion cffects has found that people
consume more food in the presence of others who are
al5o consuming (Zajonc 1968). Terkel ( 1970) provides
anecdotal e 'idence that. during the Great Depression.
there was an upswing in gambling on horses and num-
bcrs in this cxample. thc relevant reference point may
Loewenstein ( 1990) experimentally demonstrated the
effect of social comparison on impulsi vil y. Fifth- graders
who won a moderate prize ($4) in a competi- tion against
opponents who won a larger prize ($8) displayed greater
inn patience than subjects who won the same moderatc
prize but whose opponents won a smaller prize ($ 2).
Losers were willing to give up more than winners to get
the reward inn mediately (450 vs. I 2e). Losers also
requi red a larger premium than win- ners ror enduring a
further delay (534 vs. 23c). If losers use their lucky peers
as a consumption standard and partially adapt to what
lhey might have gotten if they had won, then the
reference-point m odel or desire pre- dicts greater
feelings of deprivation and impatience for losers than
for winners.
Advertising based on modeling and identificalion
with prototypical product users (Peter and Olson
1987) may induce reference-point shifts through
social com- parison. A n cxam ple is VA LS-inspired
ad yen ising (e.g., American Express) that targets
emulators, who model much of their behavior after the
more materially successful achievers (Hot man 1984).
Although iden- tification with a product user
undoubtedly increases desire by increasing the slope
of the value function on the gain side, our analysis
suggests that identification may also encourage
buying through a rightward shift of the consumer’s
reference point. Of course, the strength of the
reference-point shift will depend on the level of
identification with the product user (Krebs 1975).
SELF-CONTROL: THE CONFLICT
BETWEEN DESIRE AND
WILLPOWER
Consumers are not passive victims of their own
fluctuating preferences. Those who have experienced
rime-inconsistency and its consequences (e.g., the re-
gret associated with closets full of unworn clothes or
embarrassing product returns) are likely to develop
se!f-control strategies for imposing consistency on
their own behavior. I n this articlc. we frame the self-
coMrol problem as a psychological conflict between
desire (the initiating hedonic force previously
discussed} and willpower (strategies used to
overcome desire).
Two-factor models of behavior have been an
integral part of many accounts of self-control (Freud
191 1 ; James 1890: McIntosh 1969; Winston 1980).
In psy- choanalytic theory, the conflict is represented
as an oscillation between primary process thinking
(the id or pleasure principle), which is impulse-driven,
largely irrational, and seeks immediate gratification at
any and all costs, and secondary process thinking (ihe
ego or réality principle), which is patient. logical, and
has the will to postpone gratification in lhe ser vice of
future long-run gains or goals (Hilgard 1962).
JOURNAL OF CONSUMER RESEARCH
The ability to maintain self-control and successfully
implement long-run decisions depends on the relative
strength of the opposing forces of desire and willpower.
The consumer’s conflict is portrayed in Figure 3. The
consumer can rest at any placc on the diagram. The
x- and J'-coordinates represent current levels of desire
and willpower. In Figure 3s, the consumer begins with
a standing lcvel of desirc for a product, D . To illus-
trate, imagine a consu mer who has prcviously decided
on a long-term course of action, weight control, and
vows not to eat desserts. At D , willpower is greater
than de5ire, so the consumer does no1 want to pur-
chase. If, however. the consu mer were to come into
direct contact Uh a previously rejected alternative
(e.g., by being confronted by the dessert cart after a
lavish meal), the increased proximity might boost de-
sire and move the consumer righ tward to D› . With a
large enough increase in desire, the consumer can cross
over the “buy li ne” to ltte point where desire now
dominates willpower. At this point, there is a need for
self-control. A time-inconsistent purchase will result
u nless the co nsumer can quickly move back above the
dashed diagonal.
Figure 3b sh ows that there are several ways that
the consumer can respond. First, s/he may buy, and
pos- sibly consume, the product immediately in an
attempt to sate the desire; in such cases the co nsumer
never moves from N, . Purchases of this kind may
occur with the minimu m conscious deliberation
charactcristic of automatic or mindless behavior (
Langer, Blank, and Chanowitz 197d; Langer and I
mber 1980; Weinberg and Goltwold 1982), or the
pure reaction with little or no cognition (Holbrook et
al. 1990). Alternatively, consumers may experience an
“interrupt” (Bett man 1979; Simon 1967) that alerts
them to the need for cognitive deliberation. Interrupts
can assume many forms. Consu mcrs may recognize that
they are in a situation in which they prcviously made
purchases that were regretted (e.g., in a fancy clothing
store right after payday). Thc act of pa ying fo r the
ilcm may also act as a n interrupt. Even desirc itsel f
may serve as an in- terru pl: loo su d den an increase
in desire may lead consumers to be suspicious of their
motives.
The desire-willpower framework illustrated in Fig-
ure 3b ma kes it clear that Ihere arc two distinct forms
that self-control attempts can take: desire-based and
willpower-based. First, consumers may attempt to di-
rectly reduce desire (a direct movement to the left) by
physically or psychologically reducing proximity to the
product. Alternatively, consumers may attempt to
overpower desire (an u pward movement) by relying
on a variety of willpower strategies. Willpower at-
tern pts will be succes rul to the extent that consumers
can nor oifer persuasive rationalizations (a downward
movement) thai might legitimate a time-inconsistent
prefercnce.
Before moving on, it is important to note that con-
sumer self-control, however, need not be as reactive
CONSUMER SELF-CONTROL 49'7

FIGURE 3

DESIRE-WILLPOWER MODEL OF SELF-CONTROL

( a) TIme-inca nsistent preference induced by prox imity ( b) 8 fra legies for ma in ie in in g set f-com rot

W i I I powe r

D
o
w
nl
oa
de
d
fro
m
Desire Desire htt
ps
://
ac
as suggested in the desk re- ind uced, interrupt vie w. Although standard models of dccision making take ad
There are other rcaso ns people might want to reg ulatc e
tastes as fixed and inx'ariant, there has been some work
their inn mediate consumption hehavior. In his discus- mi
on the sclf-manipulation of preferences. Most rele vant c.
sion of modern hedo nism. Campbell ( 1957) offers a
to our rcfercnce point m odel is the literature on as- ou
m uch more proactive conception of selt-control. For
piration level. AT though aspiration lcvel sometimes is p.
Campbell, m uch pleasure is sought vja emotional co
considered as cx ogenously specified, much of the lit-
stimulation (imagination. window shopping) to create m/
erature h‹is rccognix.ed that individuals have some
a necessary level ot novelty. Drawing on earlier work jcr
abilit y to control their own aspiralions. Downward /ar
by Scitovsk y ( 1976), Campbell offers a conccption manipulations of aspirations may serve to reduce the
of self-control that could be labcled “pleasure ticl
frustration arising from failure to meet goals (Festingcr e/
manage- ment.” Consumers cngagcd in p[easurc managc 1942). Likewise, people may inflate their aspirations 17
ment recognize (at least implicitly) that there is a to increase their motivation lcvel (Frank 194 I ). In our /4/
trade-off between comt“ort and pleasure. Scitovsky framework. people manipulate lheir reference points 49
defi ned comfort as residing at an optimal state of 2/
to maintain consist ency in their own behavior. Like a 17
arousal. He argued that comfort i n itself is not surge protector on a computer, the goal of such efforts 97
pleasurable: pleasure is generated during the process of is to neulralize transient shi fts in desire that are caused 24
getting to the com- fortable state (eithcr incrcasing or by sudden increases in proximity. This scction outlines 3
decreasing arousal to thc optimal level). Pleasure threc self-control tactics (avoidance, postponement by
manage menl in vol pcs self-imposcd periods of deprivat gu
and distraction, and substitution} that reduce desire es
io n to inc re asc futurc pleasure, It demands consumer b y undoing refcrence-point shifts. Each tactic inn plies t
foresight. requiring a consumer worldly enough to a lcftward movement in Figure 3. on
appreciate George Ber- nard Shaw’s obser v ation t hat 09
“There are lwo tragcdies in life. One is not to get your Ai'‹ iJunt’e. Clearly, the best way for consumers to O
heart’s desire. The other is to gai n it.” Everyone avoid time-inconsistent behavior is to eschew situa-
engages in pleasure manage- ment to some degree. Not lions in which they are likely to cxperience increases
eat ing bet ween m cals so as to better enjoy din ner, in desire for previously rcjectcd alternatives. Avoid-
and taking a 10-kilometer run so as to experience lhe ance prccludes physical or sensory proximity and, thus,
pleasurc of slopping, are ex- amples of defcrring prevents a proximity-induced shift in a reference stan-
short-term gains for the promise of long-term dard. The importancc of" avoidance is well-known to
pleasure. Pleasure management is a fas- cinating form those who have given up alcohol or drugs (although
of sell’-control that deserves more study. there is some controversial clinical evidence that al-
coholics and addicts may be able to consume in m od-
Desire Reduction eralion: Polich, Armor. and Braiker 19S l ). Recogniz-
One way in which consumers can maintain self- ing that their resolve could lapse under the wrong
control is by manipulaling Ihcir own reference points. circumstances, recovering alcoholics and addicts
nften
500 served. Rook and Hoch (1985) found that consumers
employed a variety of distancing strategies based on
avoid bars and parties where alcohol or drugs are avoidance, such as: “You’ve got to walk away—as soon
as I feel an impulse, I immediately leave the area,”
and “I steer clear of record stores whcn I can’t afford
it.” Schifter and Ajzen ( I 9S5) found that women were
more successful in losing weight when they had de-
veloped a fairly detailed plan of action to deal with
the unanticipated difficulties of dieting.
Po.st pon‹'ment and Dislruc lion, Postponement
means putting off a choice until some ruture date, an
effective st rategy against transient desires (Fig. 2b).
Many consumgrs devise post ponement rules, such as:
“Never buy a car on the first visit to the dealership,”
or “Always consult my spouse before making a major
purchase.” After a consu mer leaves i he showroom or
department store, his or her desires, which had seemed
firm ly entrenched. often recede. Rook ( 1987) found
that the most difficult moments in a n impulsive-buyi rig
situation occur inn mediately after the i mpulse to con-
sume is first felt. Consumcrs believed that the intense
desire would subside by itself if they could only resist
it for a little while. Recognizing that impulsivity may
be cultivated by high-pressure sales t actics, nu mcrous
states have th ree- to seven-day cooling-off periods
during which a consu mer can reconsider and nullify
a purchase. Postponement, howevcr, does not work
for all people. N isbett and Kanouse ( 1968) found that,
although there was a positive relation between un-
planned inn pulse buying and number of hours of food
deprivation for normal-weight individuals, there was
no relation for obesc consumers. I n fact, obese con-
sumers purchased even mo re when they had just eaten,
suggesting the possibility of an eating-begets-ealing
mechanism.
Distraction has proved to be one of ihe most suc-
cessful means of postponing. Whereas attention to the
delayed-goal object increases the pain associated with
waiting, distraction reduces the frustration and arousal
that accompanies the typical waiting period,
effcclively reducing the magnitude of the shaded
areas in Figure
2. Work by Mischel and his colleagues has dc mon-
strated that self-distraction is an effective way tor chil-
dren to maintain self-imposed delays of desired goal
objects (other than avoidance). 1 n two sets of experi-
ments, subjects were i nstructed to distract themselves
while waiting (Mischel and Ebbesen 1970) or to cog-
niti vely transform the inn mediate rewards (e.g., by
thinking of the marsh mallows as liltle White clouds;
Mischel, Ebbesen, and Zeiss 197 2). Both distraction
and cognitive transformation increased willingness to
wait.
Subsiif ulion. Substitution involves offering oneself
a small but immediate reward for successfully resisting
a larger impulse. For instance, as a snack, a dieter may
substitute a piece of celery for the much more satisfying
CONSUMER SELF-CONTROL
choice as a problem of collective action between a se-
quence of temporally situated selves. In the same way
JOURNAL OF CONSUMER RESEARCH
pean uts or potato chips. The purpose of this tactic lfi
to provide sufficient immediate satisfaction to endure
the deprivatio n associated with waiting, perhaps si-
m ull aneously distracting oneself from the original de-
sired object. Low-calorie foods (e.g., light and nonal-
coholic beverages and high-fiber breads that fill you
up but not out) are positioned as wholesome substitutes
for the real thing. Specialty shops and galleries selling
expensi ve antiques or artwork also offer relatively in-
expensive baubles and gewgaws that can be purchased
to sale desires for more imposing objects.
The effectiveness of substitution is uncertain. Sev-
eral fi ndings from the delay-of-gratification paradigm D
suggest that it can backfi re (Yates and Mischel 1979). o
Although substitution may momentarily distract the w
consumer from the more desirable object, the delayed nl
oa
object may come back into focus. I mpatience may in- de
lensify when 1he small reward is no longer available. d
Moreover a “lesser evil” substitute may not always be fro
an improvement over the focal object of desire, as wit- m
nessed by people who continue to sm oke in order to htt
ps
avoid weight gain ( K lesges ct al. 1989). ://
ac
ad
Willpower e
The second factor influenci ng self-control is will- mi
c.
power. Willpower refers to the diverse tactics that peo- ou
ple use to overcome, rather than to reduce, lheir own p.
impatience. The idea of pvercoming desires is inimical co
to lhe rational choice perspective, which assumes thai m/
people try to satisfy, raihcr than to ovcrcome, desires. jcr
/ar
The use of willpower tactics implies the existence of ticl
multiple, quasi-independent centers of motivation e/
wilhin individuals (or m ultiple selves [Elster 1977]). 17
I n literature and the fine arts. self-conflict is often /4/
portrayed as the soul caught between the devil’s im- 49
precarious and an angel’s moral ex hortations. I ntra- 2/
17
personal conhict is a recurring theme in consumers’ 97
self-reports of impulse purchases (Rook 1957; Rook 24
and Hoch 1985). To capture the introspective expe- 3
rience of intrapersonal conflict, theorctical frameworks by
gu
of inter personal conflict have been used to model the es
struggle between multiple set ves. Ga me theory, the t
problem of collective actio n, and principal agent the- on
ory ha ve all been translated into intrapersonal terms. 09
Schelling (1978) and, more recently, A inslie ( 1985) O
view intrapersonal conflict as a repeated two-person
game, an “intimate contest for self-command”
(Schelling 1984) between two selves, one myopic and
the of her farsighted. For example, there is the self
(dominant right after smoking a cigarette) who wants
to quit smoking and the self who wants “just one
more.” There is the self who wants to rise early (and
sets the alarm clock accordingly) and the self who re-
jects the previous night’s resolve in favor of sleeping
i ri. Willpower. in this context, can be viewed as t he
efforts of the farsighted self to constrain the behavior
of the m yopic sclf. Elster (1977) views intertemporal
that social equilibrium can unwind as the result of in-
dividual defections, inn pulsi ve acts by earlier situated
selves (e.g., smoking a cigarette) can lead to an ever-
expanding series of defections. Thaler and Shefrin
( 198 I) apply a theoretical framework adopted from
principal agent theory in Eco no mice to understand in-
trapersonal conflict. In their model, an a temporal, far-
sighted planner (the principal) attempts to regulate the
behavior of temporally situated, shortsighted doers (the
agents).
Sjoberg ( 1980; Sjdberg and Jnhnson 1978) views
willpower as a form of high-quality deliberation re-
quiring cognitive effort. “To have a strong will means
being able to stick to an initial well-balanced decision
under various forms of pressure” (Sjiiberg and
Johnson 1978, p. 1 50). Volitional breakdowns occur
under the influence of strong moods or desires, when
“some en- ergy which otherwisc would have been
available for the cognitive system is lost.. . . The
withdrawal of energy first affects more sop h isticared
cognii ive mech- anisms leaving the more primitive
ones. This may leave the door open for a corrupt,
twisted, and shortsighted reasoning which generates
excuses for changing the initial decision” (p. 1 5 1).
Sjoberg and Joh nson tested their theoretical
perspective by repeatedly interviewing smokers who
attempted to quit. As predicted, reversion to smoking
gencrally occurred at ti mes of extreme stress; su
bjects who resumed smoking identified cog- nitive
distortions of reality that occurred prior to re-
sumption of smoking.
In what follows, we classify all of lhe i nterpcrsonal
strategies ihal people can apply intrapersonal ly under
the category of “willpower.” People can attempt to
regulate their own behavior (by erecting constraints
or altering incentives) in the same way that they at-
tempt to regulate others' behavior. I ust as indi idtials
may attempt to use reason and argu ment to persuade
an other perso n io adopt their pcrspecti ve, t hey may
apply analogous persuasion tactics to themselves.
Willpower tactics diffcr from desire-reduction tactics
in an important way. Willpower is a force that opposes
desire. Pure will power inn plies an u pward move in the
desire-will power model in Figure 3. Exercise of will-
power need not change the level of‘dcsire experienced
by the individual, although it can decrease, or inad-
vertently increase, the intensity of a desirc in certain
cases.
Willpower-based strategies include all attempts by
consumers to enumerate and make salient the costs of
satisfying time-inconsistent preferences. There are
economic costs, based on an assessment of the ability
to pay, and psychic costs, such as guilt and regret. We
also discuss consu mer rationalizations thai might un-
dermine the effectiveness of particular willpower
st rntegies.
302
today and the ability t o make a speci hc purchase in
the future; this may be especiall y true in households
where the spender is not ihe bill payer. Sr udies of
50 1
Per 'ornm // m en/ Precommit ment involves any
device through which consumers impose constraints
on, or n ltcr inrenti yes for, future behavior. A classic
example of precommitment was Ulysses’ instructions
lo his crew t o bind him to the mast so that he could
hear thc Sirens without jumping overboard to his
death: his crew stuffed wax in their ears so that they
could sarely sail the boat (Elster 1977). Wiring one’s
jaw shut as a means of losing weight and taking the
d rug Antabuse io discourage future drinking arc mod-
ern examples of precom mitment through self-binding,
as are ltte less extreme cases of placing the alarm clock
across the room (Schelling 1978) and leaving one’s D
c re dit cards and chec kbook at h ome when going shop- o
ping. “Side bets” (Becker 1960) are also often included w
in the precommit ment category. Side bets are contracts nl
that com mii consumers to a course of action in which oa
future rewards are irrevocably tied io the ability to
de
d
avoid more inn mediate satisfactions. fro
Precom mitment is usually assumed to operate by m
brute force. The individual either eliminates the option htt
of consu ming or imposes sue h extreme penalties on ps
impulsive behav›or ihai the costs outstrip the benefits.
://
ac
However, if credible, precommitment may also have ad
an impact th rough its effect on desire and impatience. e
Knowing that one will not have the option to consume mi
in the future can rcducc desire. For example, addicts c.
ou
surer le5s withdrawal pain when deloxifying in an es- p.
tablishment that has a reputation for incorruptibility co
(SChelling 1984). m/
Precommit mcnt has received much attention, es- jcr
/ar
pecially in recent economic analyses of self-control
(Elster 1977: Strotz 1956; Thaler and Shefrin 198 1 ).
ticl
e/
It remains an obvious enigma for the standard time 17
discounti rig view. Despite the abundance of literature /4/
on the subject. the prevalence of precommit ment in 49
aiding day-lo-day consumer self-control is unclear. 2/
Precommitment strategies are often difficult or costly
17
97
to initiate and are of limited effectiveness. Consumers 24
are adept at constructing easily revocable side bets, 3
inventing rationales for why a current purchase is a by
valid exception to the rule. Even individual retirement gu
accounts, pensions, and trusts have provisions for early
es
t
wiihd rawal. on
09
Ec'onomi Cr›sr A ’s‹'ssmeni. When consumers are O
asked what types of self-control devices they use to
resist impulsive buying, conscious consideration of the
purchase’s economic coSt5 iS often the first tactic men-
tioned (Rook and Hoch 1985). Self-control based on
cost-benefit considerations is not always easy. While
desire is proximate and vivid, economic consequences
usually are remote and di fficuli to define (Hirschman
1977). Ten dollars spent now results in a minute de-
crease in savings, or an insignificant increase in dcbt,
and is quickly lost in the complex flow of income and
purchases (Joh n SOfl, Kotlikoff, and Samuelson 1987).
There is little relation between an isolated expenditure
se[f- heI p groups for compulsive shoppers (e.g.,
Spender- tenders; see O’G ui n n and Faber 1959) have
Iound that, when mak ing purchases, profligate
spenders sometimes relied on a “just-charge-it” logic:
appar- ently, they were are able to succcssfull y
 divorce the pleasures of purchase from the economic
realii ies of paymc nt.
Consumers may make numerous rationalizations
(some provided by salespeople who have read personal
sell ing man ual s) t hal sabotage their attem pts at ra-
tional cost assessment: most notable are feelings of
entitlement. Entitlement represents t he quintessential
Yuppie theme—“you work hard, so you deserve lo
play hard.” As the theme was expressed on a recent bus-
stop advertisement for women’s designer clothes, “Beiny
it all is hard work; havink it all makes u p for it.”
Splurging is acceptable because a penance has been paid
in advance or will be in the near future. The V
ALS-inspired achiever ads rely on the entitlement theme.
as exemplified by a print ad sh o wing two ex- hausted
racquet ball players as backd rop to the copy: “He works
as hard as hc plays. And he drinks Johnnie Walker.” A
recurrent lheme in consumers’ discussions of why they
engage in recreational shopping and im- pulse buying is
that it relieves feelings of depression (Bellenger and
Korgaonkar 1980). Entitlement offers t he consu mer
the argument that “you do not deserve to feel
miserable; buy it and cheer yoursc If u p.” En- iitlement
sometimes is supplemented with a “greed- is-good”
rationale. During the G ildcd Age ( I 860- 1900), some
social commentators argued that con- spicuous
consumption offered positive benefits by en- couraging
seir-help behavior in thc middle class, a class that was
fueled by the social emulation motive (Muson 1951 ).
Entitlcmcnt may provide on ex’en more potent ra-
tionalization for purchasc when it is coupled with so-
cial Com parison, w hich, as mentioned earlier, can
cre- ate impatience by increasing proximity to the
product. The basic argument here is that, not only do
you de- serve to be compensated ror your good deeds
or hard- ships, but, in addition, since “everybody else
is doing it, you ought to jump on the bandwagon as
well. This attitude would result in a shift toward thc
bottom right- hand corner of Figurc 3fi. Since the
movement is per- pendicular to the buy line, such
rationalizai ions may
dramatically increase impai ience and thc propensity
for time-inconsistent behavior.
Time Binding. In addition to thinking about the
negative consequences of consuming, consumers may
also toe us on I he positive benefits of delay, a tactic
known as “time binding” (Jones and Gerard 1967).
Freud ( 191 1) believed that the ability to cathect images
of desired but delayed gratifications was necessary for
CONSUMER SELF-CONTROL
eating of future cones, lhe consumer may view the costs
of the cone not as 250 calories, but as an extra 250
JOURNAL OF CONSUMER RESEARCH
I he deve1 opment cif frustra4 ion (deprivation)
tolcrance. Hence, he viewed idcalion about delayed
rewards as, in cifert, a substitution tactic where i he
intermediate rewards are psychic rather lhan material.
Time binding and other typcs of internal reinforcers
have becn the subject of m uch study, particularly in
developmental research on delay of gratification (see
Mischel [1974] for a sum mary). When waiting for a
gnal object is involuntary (i.e., an extcrnall y imposed
delay thal cannot be overcome), time binding has been
dem onstrated to be an effective st ratcgy to reduce
frustration (M iller and K.arniol 1976). However, when
the situati on in vo1x•es a self-i m posed delay of gratih- D
cation (arguat›1y the most com mon situation in con- o
su mer behavior), experimental research overwhelm- w
nl
ingly dem onstrates that time binding does not work oa
well. Mischel et al. ( 1972) found that instructing sub- de
jects to thin k about delayed rewards while waiting for d
them made subjects less willing to delay gratification. fro
A n increasing unwillingness to delay gratification is m
htt
especially like1 y when delayed-reward objects are sim- ps
ilar to immediate-reward objects (e.g., both are con- ://
sum m at o ry objects; Misc hcl and Baker 1975). ac
What accou rrts for ltte limited efi’ectiveness of ad
time binding (as opposed to pleasure management, e
mi
which docs seem to be effective)? One reason may be c.
that virtually all delay-of-gratification research has ou
used child subjects. a population with limited. if any, p.
plca- surc-managc ment intuition or st ills. A n other co
possi- bility is ihat, even though focusing on the m/
delayed ob- ject makes the benefits associated wilh jcr
/ar
waiting more salient, it may also inad vcrtently ticl
increase proximity. which, in t urn, activates des ire e/
and incrcases inn pa- tience. Thin king about how 17
good dessert will taste to avoid overcal ing appct izers /4/
may actually increase cur- rent hunger pangs. thereby 49
2/
increasing the likelihood of‘ time-inconsistent 17
consumption (R uderman 1986). However. t ime 97
binding may be more effective in prac- lice than is 24
indicatcd by empirical research. Time binding has 3
been sl udied in situations in which subjects have had by
gu
to rcact to a desirable stimulus. It may be that time es
binding docs not work as well as a reactive self- t
control strategy in rcspondi rig to sudden increases in on
desire, even though it does work as a proactive form 09
of‘ pleasure management. O
Bundling ‹¿/’ C’n sis. Another tactic for increasing
the salience of purchase costs is to bundle a series of
otherwise isolated actions (A inslie 1975). Through
bundling. impulsi ve transgressions become inseparable
(as in a formal budget); thc influence of the transgres-
sions is felt in other areas, such as thc postponement
of more important and needed purchases. The old ad-
age “calories add up” is an explicit form of bundling.
Rather than myopically view lhe eating of an ice-cream
cone as an isolated act, a dieter may attempt to reframe
it as the hrsl in an endless string of self-control vio-
lations. By bundling eating of the current cone with
calories a day for t he foreseeable future, with obesily as
the inevitable outcome. to aid consu mcrs in bu n- dling,
public-ser vice television ads show piled-up car- tons of
cigarettes with the yearly cost of smoking pull- ing on
center-screen. Similarl y, the E nviro n mental
Protection Agency requires automobile and consu mer
appliance manufacturers to display a sticker 5t £ttiRg
average yearly fuel costs. Apparen ily having these cx-
penses exprcssed as a lum p su m mak cs them more real
and consequential to consumers’ future well-being.
It should be noted that actions can be u nbundled as
well as bundled. I nslead of bii nd I ing i m puls ive acts,
consumers can just as easily treat such episodes as iso-
lated even is, rationalizing their behavior with logic
such as “just one won"t hurt” or “a liitle is nor that
bad for you.” Mar kelers, also, may reframe the eco-
nomic cost of a purchase into smal lcr, more palatable
units. “Pennies a day” represents onc example of such
reframing through unbu nd ling. Cred it-card broch urcs
that state the cost of maintaini rig a deirit ba lance as
the mini m u m monthly payment, rather t han as I he
interest cost, similarly focus attention on the small size
of payments: these brochures di verl attention from the
more distant problem of how long the payments will
last. Consumers ma y also attem pt to redefi nc purchases
so that they appear time-co nsistcnt. Sales promotions
can provide a basis for rationalization by raising the
spectcr that it would be shortsightcd not to make the
purchase. A food firm’s recent direct- mail offc r shows
a large full-color picture a n fanc y dessert with the
headline, “Would you throw away a ?"R EN Dobosh
Torte worth $1 1.95?” The offer increases desire by
shifting the consumer’s reference point (i.e., not or-
dering results in a loss) and also pro• ides a means for
rationalizing other catalog purchases. Realtors are wcll-
versed in highlighting the u niquc characteristics of a
ho use and warning that if one does not make an offer “it
may get snapped up.”
Hi pher N of/i riiJ'. Consumers may also invoke higher-
orde r principles, or religious doctrine, in their efforts to
resist impulsivity. Appealing to higher au- thority is
similar to cost bund ling in that the costs as- sociated with
succumbing to a momentary desire arc treated as global in
nature. Violations that could be considered as not hing mo
re tha n a sin glc lapse are re- cast as significant
transgressions aga inst ‹inc’s beliefs and values. Negative
consequc nccs cannot be local- ized. The ba8tS of self-
control in the higher-authority situation involves strict
adherence to rules and regu- lations. Any momentary
violation, no matter how small in isolation, represents a
serious breach.
Higher authorities take a number of forms, both in-
ternal and external. The consu mcr may pray for
strength, rely on peer grou ps, or treai the tempting of
self-control as a challenge. When there are no
external
504
in desire. Specifically, it shows how a nonproblematic
standing desire can turn into one that fosters time-
503
authorities available, higher a ulhority can take the
form of tal king to oneself or related forms of self-rc-
in forcement (Mischel ct al. 1972). Relying on higher
authority. however, can be a risky strategy. When
transgressions occur, as is likely to happen occasion-
ally, instead of continuing to struggle for self-control,
individuals may view themselves as lost causes and
abandon all further restraint. Research on dieting
(Loewc 1982) has found that binging behavior is
often precipitated by reference to an “already-blown-
it” logic.
Rezr‹•t arid G mill. I n additio n to the material costs
associated with time-inconsistent purchases, there are D
o
also psychic costs (O"Guinn and Faber 1989). Research w
on decision making under uncertainty (e.g., All ais’s nl
paradox) has found that people avoid actions that they oa
anticipate regretting (Bell 1982). Consumers may also de
anticipate future gu ilt and shame. I m pulsive purchase d
fro
behavior may conflict with basic mo rals and val ues,
m
signaling hedonism, materialism. or selfishness (Belk htt
1982. 1985). Consumers may worry about the con- ps
spicuousncss (Mason 198 I ) or fear the embarrassment ://
thal may accompa n y scc mingly frivolous purchases. ac
The embarrassment need not be public, as witnessed ad
e
by consumers’ feel ings of immaturity when looking mi
into a closer fi lled wilh infrequently worn shoes and c.
clothing. ou
A rccc nt automobile ad, recognizing that guilt can p.
co
reduce the probability of purchase, offered explicit m/
countervailing arguments. lt read, “To anyone who jcr
thin ks a Mercedes-Benz S-Class is self-indulgent, a /ar
brief lesson in self-prescrvai ion .” A detailed pictorial ticl
e/
and verbal enumeration of the product’s safety features
17
followed. The ad ended with the tag line, “the only
/4/
luxury sedan that can offer the luxury of being rated 49
‘the safest car in A me rica’ two ycars running.” Finally, 2/
i n many cases, recognition of futurc psychic cosis may 17
97
diminish desire in the short run. For example, antic-
24
ipuled rcgre Is about not practicing safe sex may directly
decrease one’s current desire. Willpower-inducing guilt
3
by
accompanied by desire-reducing fear may be one of gu
the most effective self-control devices since it tnoves es
thc consumer away from the buy line (perpendicular t
toward thc upper left-hand corner of Figurc 3l as rap- on
idly as possible. 09
O
CONCLUSIONS
I n this article we have attempted to explain how
consumers maintain self-control in the face of time-
inconsistent preferenccs. We have conceptualized
consumer self-control as a struggle between the two
psychological forces of desire and willpower.
Our desire-willpower model borrows heavily from
previous two-factor theories or self-control.
However, the model is also unique in several regards.
Firsl, using
a decision-lheorelic analysis of reference points, it
pr vides an explicit mechanism for sudden increases
inconsistent behavior. Second, whereas previous two-
factor models have tended to 1 u m p all self-control
at- tempts into one broad category, wc are able to draw
clear distinctions between desi re-based, will power-
 based, and combination stralegies. In particular, the
model recognizes that self-control is influenced not
only by environment at factors that in0uence
proximity and desire but also by the ability of
consumers to ex- ecute their own desk re- and
willpower-hased strategies. The desire-willpower
framework offers testable pre- dictions concerning
time-i nconsistcnc y a nd self-con- trol, many of which
are supported by the existing lit- erature. Additional
iests of the model could in volve increasing desire by
using proximity to induce refer- ence-poini shifis, and
ihen examining the effects of the different desire- a nd
willpowc r-based strategies on subjects’ ability to
delay consumption. Self-control strategies could be
provided through explicit instruc- ti on. It won Id also
be interest ing to in vestigale ihe ef- fects of different
forms of rationalization on self-con- trol; however.
this type of research is not easy to do. Al mosi all of
the work in the delay-of-gratification paradigm
(Mischel 1974) has used very young subjects who
have not been well socialized to the need for self-
control. On the other hand, most adult consumers al-
ready have well-developed repertoires of seir-control
tactics, especially when they are placed in situations
in which they know that they are being evaluated.
Therefore. adults may not require experimenter in-
structions to maintain self-control.
lt would also be inieresling to examine how self-
control evolves over a family life cycle in which dis-
posable-income levels and family expc nd it ures are
typically negatively correlated. Because of m ultiple
decision makers, the study of self-control would in-
volve both intra- and interpersonal con fi ict. Savings
behavior is another i mportant research application.
Saving is the mi rror inn age of consumption. Although
it has rcceivcd plcnty of attention from economists,
savings behavior has been neglected by consu mer re-
searchers (cf, Katona 1975). For reasons that are not
understood by economists, the U niled States currently
has one of the lowest savings rates among ind ustrial-
ized countries. Since consumer self-control underlies
the national savings rate (Shefrin and Thaler 1988), a
better understanding of consu mer set f-control at the
micro level could increase our understanding of
saving and its determinants.
Our work also bears on impulsc-buying behavior.
Although the term “impulse buying” has been around
for over forty years. few theoretical or empirical ad-
vances have been made. Parr of the problem is that
people have used a single term. “impulse buying,” to
refer to a wide range o(processes that cause
consumers to make unplanned purchases. Global
terminology may obscure m ore than it ill uminates.
In t ry ing to dis-
CONSUMER SELF-CONTROL
in 'Eli ‹• .Minh iylr S‹•lf. ed. Jon E lsicr, Cam hridgc: Cam-
JOURNAL OF CONSUMER RESEARCH
tinguish between the different motives for unplanned
purchasing. Stern ( 1962) developed four conceptually
distinct categories of inn pulse buying: reminder, sug-
geslion, plan ned, and pure. It is clear that not all im-
pulse buying represents ti me-inconsistent be havior,
but lime-inconsistency and consumer self-control are
closely linked to “truly impulsive buying, the novelty
or escape purchases which break the normal buying
pattern” (Slern 1962). Our approach is compatible
with the narrower focus of Rook ( 1987, p. 19 I ). who
defi ned tmp ulse buying as occurring “when a con-
su mer experiences a sudden, often powerful and per-
sistent urge to buy sonnet hing inn mediatel y. The im-
pulse Io buy is hedonically complex and may stimulate
emotional conflict. Also, i mpulsc buying is prone to
occur with diminished regard for its consequences.”
We feel that the bcst way to make progress in under-
standing impulse buying is to be specific about the be-
havior in question. Our framework may help in this
regard.
O ur econom ic-psy chological model of con sumer
self-control stands at the intersection of two broad
current s in consu mer behavior research. One
perspcc- tive views decision making as ralional and
dispassion- ate; the other views it as visceral and
emotional. Th us, the desire-wit 1 power framework
provides an ideal ar- ena for examining the interaction
between rational and hedonic motives. These two
types of psychological processes are normally com
part mentalized into sep- arate literal ures. Although
each perspective adequately describes a side range of
consu mer behaviors, neither alone can provide an
adequate account of the entirc decision-making
process. Emotional factors are re- flected in the
reference-point model of deprivation and desire.
Cognitive factors are reflected in the delibera- tion
and self-control strategies that consumers utilize. A c
hange in ciI her desi re or willpower can cause the
consumer to s hift over the buy line, resulting in a pur-
chase. Moreover. we have cited numerous cases in
w hich emotions in 0uence cognit ivc factors (e.g., desire
motivating a rationalization of the negative consc-
quinces of a purchase) and vice versa (e.g., cost anal-
ysis reducing a desire). Although concepiually
dislinci, the psychological faclors of desire and
willpower are by no means independent of one
another.
REFERENCES
Abelson. Roberl P. (1963), “Computer Sim ulatio n of ‘ £tot’
Cognition.” in C’vinpti Ier Sirn ulation r› Persr»ta/ifj', ed.
S. Tom kins and S. Messick, New York: Wiley.
A inslie, George (1975}. “Specious Reward: A Behavioral
Theo ry of I m pulsi veness and I mpulsc Control.” Ps y-
eah o Ing i‹ u1 Bit llel in, 82 (4), 46.3-509.
( 1985). “Beyond M icrocconomics: Conflict among
I nlerests in a M ultipl c Self as a Determinant of Value,”
bridge U nivarsity Press.
and Varda Haendel ( 1983). “The Motives of W ill,” in
if J‹›/pg/r A s per I s ri/ z1/ci›fir›/ o nd n my; .1 h ti i c, cd. E.
Goti heil ct at., Springfield. I L: Thomas, 1 19- 1.39.
A msel, A bram ( 1958), “The Role of L rustrati›’e Non reward
in N on con lin uous Reward Sit uations,” #i r ho/‹izi‹ al
Bull i i n, 55 (2), 102 — 1.19.
( 19621, “Frustrativc N onreward in Partial Rcin-
fc›rcement and Discrimination Lear ning: Some Recc nt
H istory and a Theoretical Fxieusr on,” Ps;‹’h‹›/‹izi‹ a/
R‹• 'i‹ , 69 (4). 306—3 28.
A us tin. Wi lli am ( 19 7 7), “Equity Tlicor y and Social Com-
parison Processes.” in S‹›c'/o/ L'‹›mpori.r‹›n 6r‹›rr.‹z‹’ . ed.
Jerry M. Suls and R icha‹d L. Miller, Washingt‹›n, DC:
Hemisphere, 279—306.
Bass. Fran k M. ( 1 V74), “The Stochastic Thcoty or PrcLer-
ertces and Brand S itc hing.” J‹›liriial ‹›f,\furkt’/ind R '-
.›'‹°orc/7, 1.1 (Fghruary), I —20.
Becker, Howard S. ( I 960)„ “N‹›tes ‹›n Ihe Concept of Com-
mitment,” J i»c’ri‹an ./‹›ui GUI ‹›I .Y‹ 6/’‹›/‹›y , 6d, 3 2—4fi.
Belk, Russell W. ( 1984}, “Cultural and Historical Differences
in Concepts of Sel I and T heir Effects on A itiludes ion ard
Having and Giv i ng.” in .4 ‹fi n ni i's in fi’run i iim‹'r fi r -
st•ur‹‘h, Vol. 1.1 . ed. Thomas Min near. Prox'o, U I’: As-
sociation for Consumer Research, 753-7f›0.
t 1985), “Material ism: Trait A spcct s of Livi ng in a
M atcrial World,” Joui’nnl rJ L.‹›n. ii m‹•i H ‹'.i thu rt h. 1. 2
(December). 2d 5 —28tL
Bell. David F. ( 19521, “Regret in C›ccision Ma king under
U nccrtai ntj.” 63 yrrul ion.s H e.s ‹•or‹’1i, 3G S ). 96.1 —9S 1 .
( 1983), “DiSappoiM ment in Decision Makin g u nder
U nccrtainty,” D¿ac'ru/i‹7ii.› A e.irai“‹'ñ, 3? t I ), I -2 7.
Bellcnger. Danny M. find P, R. Korgaonk;tr ( 19SO}. “Profiling
the Recreational Shopper,” J‹›time1 ‹›f! R‹'toilin p, S6 ( 3).
77 —92.
8enz ion. Uri, A mnon R;ipo port. and Joscp h Yagit ( 1989).
“Discount Rates 1nté rrcd Decisions: A n E hperi-
mental Stud v,” .lfzinoyeiricof .'›r i «oz r. 3.5 ( 3), 2 70—2
h4. Betl ma n, I amcs R. ( 1979 ), I n Sri/‹try glint ?’rr›‹ e.‘.i trip
1’h‹•- rr,r oJ‘ C‹›nii‹m r C/i i‹'‹', Reading. MA: Addison - WCS-
ley.
Bohm- Bawer k. Eugcn vo n I [1 h98 I ** )- ‘*F!!*! and In-
/rrr•it, Vol. 2, South Holland. I L. Libcrtaria n .
Bric k ma n, Phi I I ip and Run nic Jan of flu 1 man ( 1977).
“Pleasure and Pain in Social Comparison,” in Macie I
C‹›m ›o ri. on Nrc›r-c•.i'sc i, ed. Jerry M. Suls and Ric hard
L. M iller, Washington , DC: H emisphcre. 149- I hñ.
, Da n Coates. a nd Ro n nie Ja noff-B u1man ( 1978).
“lottery Wi n ners and A cci dent V ici i m s: Is 13appincss
Relative?” ,la urita I r›J I’ers’‹›oo/i/ r puff Arm /‹Jf Nsj’c/ii›/‹ip
I’ 36 (3), 917-927.
Campbell, Cal in ( 19b 7). The fi‹›r›to›inc ñ’i/ii‹ motel //i r .Sp/r//
‹›.f.modern L"‹ n.s iimt r/.v n, Lo ndon: Basil Bl ac k ell.
Chung, Shin- Ho and R ic hard J. H cr nstci n ( I 9fi*), “Choice
and Delay of Rein furcement.” Journo 1 of! tlw F..‘y‹ ri -
Duesenberry, James ( 19 49). int»rir, ,Sn ring. onf/ the /'ñ‹'‹»‘r
of €"‹ n.‹timc'r B‹ yu vioi’, t“s mbridge. MA: Har ard
Uni- versity Press.
Easterl in, R.A. ( 1974), “Does E con omic Grow'ih ] mpro ve
the Human Lot? Some Empirical Lvidence,” in Nai ion›
506
Klesges, Robert C., Andrew W. Meyers, Lisa M. K lesges,
and Marie E. La Vasque ( 1989). “Smo king, Body
505
‹›l .I to.s‹'.s . I o n‹›i ii., cd. P.A. David and M.W. Reder,
New York: Academic Press.
Elster. J on ( 1977). é. I i-.i v‹•. unJ flu.• Sir‹•n.s, New York: Cam-
bridge Min iversily Press.
Fcsti nger, Leon { 1942 ). “Wish, Expecl at ion a nd G ro up
Standards as L aclors In flue ncing Le vel of A spi ratio n ,”
mourns1 r//’. 1JiJfr‹ d .Sii‹‘iof P.i r‹lioloy ', 37, 1 54—200.
( 1954). “A Tlieor y of Social Comparison Processes.” I-
f i«nun £ t•lu I i on.\. 7 (2). 1.1.7 — 1 0.
Fishbu rn, Pelc r ( 1977), “Mean• Risk Analysis with Risk As-
soci:ited wilh Below-Target Retu rns.” Am eri‹ un fi‹'‹›-
ni›in /r’ fii'i i«t , b7 (2 ), I I 6— 1.26.
Frau k, J. D ( 194.1 ), “Reccnl Stud ics of the Level of Aspi-
raii‹sn,” Ns I r /i ›/‹tpit'u/ Ali//‹•f in, 38. 2.1 h—2 26.
D
Free d, Sigm un d I [ 19.1.1 ] 1959), “For m ulalions rcgarding
o
the Two Princi pies of Mc ntal Fu ncliout rig,” in C w
‹›f/‹’c'feff f'o/›c'rs. Vol. 4, New York: Basi c, I 3-2 1. nl
G urr. Ted R. ( 1970), 11 /i r tf‹'ii ft‹'ñe/, Princeton, NJ: oa
I•rinccton U nivarsity P rose. de
Ciu yton, Arth ur C. [ 197.1 ), fo,‹t‹’ fiiiiriun Uh v. uil‹ig . Normul d
I“iiii‹ init ‹i /iJ .Wr‹'/we i tin ‹ ‹›/ D‹'f‹ ns‹'. Phi ladcl phi a: fro
Sanders. m
H elson. Harrv ( 1964). I dop/arii›n-/ ‹ v‹'/ 7’/ie‹›rJ', New York: htt
Har per & Row. ps
14 ilga rd. E rncsl R. ( 1962), “I mpu Jsive versus Realistic ://
ac
Thin kin g: A n Examination of lhe Distinct ion between
ad
Primary and Secondary Processes in Though t,” N.sJ -
e
-//‹›f it/‹’ ! flir//z•frn, 5.9 (6), 477—48 8. mi
ti irsch man, Albert C›. ( 1977), ’f’/i‹ Nuss'i‹›n,s a nd the c.
in/‹•r‹•‹f,c P clit! ‹’o1.1 ryiiinr•n i ; /r»- (“npiyzifism ou
hc•(nrc• It s 1‘riwnyh, P ri nccion. NJ Princeton University p.
Press. co
H i rscli man. EU zube I h C’. and Morr is B. H olbroo k ( 19S21, m/
“H cdonic Con su m ption: E m crgi ng Cone cpls. M el hods jcr
and l'ropositi on s.” Ji›wiio I of .ilarL ct ing, 46 (Sum mcr), '? /ar
2 — I (11 . ticl
Holbroo k. Morris B.. John O’Shaughncssy, and Stephen Bell e/
17
t 19.90 . “Actions and Rcaci ions in lhc Consumption
/4/
E x pencncc: Thc Complc mcntar y Roles of R casons and
49
Emotio ns in €“onsu mcr Behavior,” in R r›'‹ ore I in C“on- 2/
.s mini i- d‹•Ji i rir›r. cd. E li zabct h C. H irschman, G ree n- 17
wich. CT: JA1. 97
Idol man. Rebecca 11. 1.1984 J. “A Val ucs and Li fcsi yles Per- 24
spccli ve on 11 u m 0 n Be havior.” i n P‹•rs vnul thu lut•.s 3
an d C no.s iim‹ i P.› r‹ b‹›/r›pj , ed. R . E. Pitis. J r., and A. R by
. Wot›d st de, Le x ington . MA: Lcxi ngton . gu
Irw i n, From cis W„ L an nic M . Ar mitl. and Charles W. Si m es
on ( 19 4 3). “S tud ies i n Object-Prefcrc nccs. I. The E ffcct t
of Fenn pural Pro ximit v.” J‹›iirn ul of £’. /ic'rimc•a/o/ P.s on
t'- r J‹if‹ jir. 3.3 ( 1 ). fi4-7 2. 09
O
Jam es. Will iam ( I h90). Tito’ Prin‹'i/›l '.‹ ‹›I /'.‹i‘‹ h‹›In¿i-, New
Yt›zk: I I r›lt, k inchart & Winston.
Juh nson. Stephen. Lawn ncc K‹›tlikofL, and W i1tiam Sam-
uelson ( 1987). “tun PcopIr Campuse* An E xperimcntal
Tesl of‘the Lily Cyclc Consumption Model.” paper pre-
sen led ut I he N B E R Research Program on Economic
FI uct uai to us.
J ones. kd• ard and Harcild B. Gerard ( 1967), “rn‹ndo/ir›n.s
of .Eni io/ /’.› i c /i«/‹›yr, New York: Wiley.
Kah nema n, Daniel and A « ’r ersk y ( 197S), “Prospect
Theory: A n Anal ysis of Decision under Risk,” ñ‹’r›n‹i-
lii i ii’/r o, 47 (2). 36.3 — 39.1.
KoiC R’o. GcolgC I *115 1. Ps)'‹’hologic’ul Emu vm res, A
mster- dam: E hex ivr.
Weighi, and Their Effects on Sm oking Behavior: A Com
prrhensi ve Review of the Literature,” I’s y‹-li‹›lop- i‹-al
Bullet in, 106 (2), 204—23 1.
Knetsch , Jack L., Richard Thaler, and Daniel Kah neman
( t 987), “Reluctance to Trade: An Experimental Refu-
 tation of ihe Coase Theorem,” working paper, Graduate
5chool of Business, Corne11 D niversity, Ii haca, NY,
14853.
Kotler, Phillip ( 1973- 197 4), “Atmospherics as a
Marketing
Tool,” Journal of Rrt ai li rig, 49 (Winter), 4fl—64.
Krebs, Dennis ( 1973), “Empathy and Altruism,” Journs1 rif
Perrys‹i lit y and .T‹›‹ ml I’s vch oMg y. 32 (6). 1 134— 1 1
46.
K y dla nd, Finn and Ed ward P rescoti (1977), “R ulus rather
than Discrei ion: The I nconsistency of Optimal Plans,”
Joiirnttl uf Polil incl I.°ci›n‹›rn y, 85, 473—49 2.
Langer. Ellen, A . Blank, and B. Chan owilz ( 19 7g), “T he
Mind less ness of Ostensibly Thoughtful Action: The
Role of Placebic I n formation in 1ntcrncrsonal I nier-
action.” Jr›«rnnf ‹aJ f'cr.sonafi/J' onJ So‹‘inl
N.tJ'c0o/r›gJ', 36, 6 35-642.
and L. I mber ( 1980). “The Role of M indlessness in
the Perception of Deviance,” Joii ryal vf Pvr.s‘‹›nalii y
and Svci‹il P› y‹holog y. 38, 360—36 7
Lewin, Kurt, Tamara Dembo, Leon Fcslinger, and Pauline
Sears ( 1944), “Level s of’ Aspi rai ion.” in Per.trans/ifj'
und Belui vi'or Dis‘order.s, ed. J. Me V. i-I unl, New
York: Ron- ald. 333—37 g.
Loewe. M.G. ( I 9fl 2), “The Rolc of Anticipated Deprivation
in Overeating.” A Edit iiv‹• Bc•ho vi mrs. 7, 10.3— 1 1 2,
L oewenslein, George F. ( l988), ‘“Frames uf Mind in I nter-
temporal Choice,” Manug mvnt S‹’ic'n‹’‹'. 34 (2), 200—
214.
( 1990), “Reference Points in 1ntertemporal Choice,”
wo rk ing paper, Center for Decision Research, Univer-
sity of Chicago, Chicago. I L 606 37.
Macl n nis, Deborah J. and Linda t.. Price (1987), “The Role
of [ magery in I n formation Processing: Review and Ex-
tensions,” Joti rnul /C‘‹›a› wn vr Re earn h, 1 3 (March),
473-49 l .
Marlutl, G. A lan, Joh n S. Bacr, Dene is M. Don ovan. and
Daniel R. K ivla han ( 198 S). “Addict ivc Be haviors:
Etiology and Trcatmc nt,” Ann ual R rrf‹'iv' ( P. y‹’Iiolvg
'. 39, 2 23—25 2.
Marshall, Alfred ( 1 S90). Print idle.i ‹›f‘ f?coiiomic‘.s, 1st
ed., London: Macmillan.
( 1920), Prin‹’iples oJ‘ I.‹’unomi‹ .i, 8th ed., Lo ndon:
Macmillan.
Mason, Roger S. ( 198 I ), Consyi‹’uous C’‹›n.s eruption. A
ited y D/‘ ñ.Té’ep/ lunol €’‹ n sum •r £te/i avi‹›r ,
Westmead, England: Gower.
McI ntosh, Donal d ( 1969), The I’oun Lai i‹ins ‹›J‘/ftimas So-
c'ie/y, Chicago: University of Chicago Press.
Merton, Robert K. ( 1968), .Arm/oil Theor y and Serra I St
ruC-
f are, New York: M acmillan.
Miller, Dale T. and Rachel Kar mol ( 1976), “The Role of
Rewards in Eeternally and Self-imposed Delay of Grat-
ification,” Jutirnal i›] Per.s vnalil y• and So‘ial P.
(ycholog v. 33 t5), 594—600.
Miller. George, Eugene Galanter. and Karl Pribrarn (1960),
Plans and ih‹ Sr ruck urt t›j Beha›'i it r, New Yo rk: Hott,
Rinehart & Winston.
CONSUMER SELF—CONTROL
Simon, Herbert A. ( 1967), “Mai ivatinnal and Emoi ional
JOURNAL OF CONSUMER RESEARCH
Mischel, Walter ( 1974), “Processes in Delay of Gralitica-
lion,” i n Add anr ve in I.’xyerim mm 1 Si›cui1 Psyt’holog
y, VoL 7, ed. D. Berkow itz. New York: Academic. 249—
292.
and Na ney Baker ( 1975), “Cognitive Appraisals and
Transformations in Delay Behavior,” .h›urnal vf Pcr-
sunnfiiJ' und Son ial Ps5 ’liolog y. 31 (2). 254-26 1.
and Ebhe B. Ebbesen t 1970), “Attention in Delay of
G rat i ficat ion,” Joiir nal o( Pier.en nulit y and Soria I
Ps y- chilled›', 16 (2), 329— 337.
, Ebbe B. Ebbesen, and A ntonette R. Zciss ( 1972),
“Cognitiv c a nd Attc ntional Mechanisms i n Delay of
Gralificat ion,” Jout nal r›/ P‹ re ronfiiJ' onal .Socu’al
1•.‹ - ’h‹›loy; . 2 1 (2), 204—2 18. D
and Joa n Grusec ( 19 67), “Waiting for Rewards a nd o
Punish menls: Effects of Time and Probability on w
Cho’ice,” Journa/ o] Prrsunali t y and Sucial Psycholoy nl
v, 5 ( I ), 24-3 I . oa
Nisbctl, Richard E, as d David E. Kano use ( 1968), “Obesity. de
H unger, and Supermarket Shop ping Behavior.” I’ro- d
i « d/rips, American Psychological Association A n nual fro
Con mention. m
htt
No rd. Walter R. and J. Paul Peter ( I 9d0), “A Behavior ps
Modificali on Perspective on Marketing,” Journal ‹›f
://
Market in g, 44 (Spring). 36—47, ac
O’Gti in n. Thomas C. and Ronald J. Faber ( 1989), “Com- ad
pul sive Buy i ng: A Phc nome n ological Explorati on ,” e
Joii rnul ‹;/' Con.vurn‹•r R‹•.s‹•u rch. 1 6 (Scptem ber), 1 mi
47 — 157. c.
Peter, J. Paul and Jerry C“. Olson ( 1987), £: on.burner Zrcha ou
vier.’ p.
co
.MarLc'i ing .Siralc•{• y Per.sp‹ cii ver. , H omewood, IL:
m/
Irwin. Polich, J.M., D.J. A rm or, and H.B. Braiker 1198 I ), jcr
Th • C‹»‹r ‹' o/ Il‹ cm›/i.inn. f’‹›or Scar.‹ ofrr /?‹•nrrnrni, New /ar
York! Wilry. ticl
Rook. Den nis W. ( 19 fl 7). “The Buying I mpulse.” lv u rnal e/
vf L.’on›‘um er Resettrt’li, 1 4 (September), l 89— 199. 17
and Stephen J. Hoch ( 19SS), “Consuming Impulses,” /4/
49
in ldrnncrs in €’onsumt•r Re. ‹•or‹/i, Vol. 12, ed. Eliz-
abeth C. H i rsch man and Morris B. Holbrook, Pro vo, 2/
17
U T: Associai ion for Consumer Research, 23—27.
97
R uderman, Alice J. ( 1986). “dietary Restra in1: A Ther›rct- 24
Ical and Empirical Review," P.sy‹hol‹iyi‹:al Rtillel in. 99 3
(2). 247—262. by
Sc helling. Thomas ( 19 78). “Egonomics. or lhc Art nf Self- gu
Management,” km‹grit un Err›nomic Revit•i‹ I'apers and es
Prof edings, b8 (May). 1 34— 1 39. t
( 1984), “The 1 climate Contest for ski r-command,” on
in C’h‹ii‹e and Con.«•qtit'nc‹•, Cambridge, MA: Harvard 09
O
U niversit y Press, 57 —82.
Schit‘ter, Debora h E. and Icek Ajzen ( 1985), “Intention,
Perceived Control. and Weight Loss: An Application of
the Theory of Planned Behavior,” Journal of Per.ii›nalii ¿
and Soci‘nl P.› y‹'liolozv. 49 (3), 843-85 I .
Scitov sk y, Tibor [ 1976), Thr Jo yle.s.s ñc ‹ni‹imJ. H n Ing
ui r y rule Huma n Satis Luc’l ion, New York: Oxford
University Press.
Shefrin. Hersh M. and Richard H. Thaler ( 1988), “The Be-
havioral Life-Cycle H ypothcsis,” /icr›n‹»x/t‘ Inpuirt.. 26.
609-643.
5il verstein, B. ( 1976), “A n Addiction Explanation of Ciga-
rette- induced Relaxation,” unpublished dissertation.
Columbia University, New York. NY 10027.
Controls of Cognition,” N.iJ'r'J‹›fng/‹.'of ftevfew, 74, 23— 39.
Sjoberg, Len nari ( 1 9d 0). “V o1it ional P roblc ms in Carry ing
Through a Di fficull Occision.” A‹’l o 1’.‹ !cholo¿i‹:a, 43, I
 23-l 32. 507
and Tommy Joh nso n ( 19 78), “Trying lo Give Up
Smoking: A Study of V‹ilitional Breakdow ns,” Addr c live
B haiii›r.s', 3. 149- 1.6.4, Consum c r Choicc,” It›urnal oj Econom i‹’ Beha•'ior ond
Skin ncr, B.F. ( 19 SG). “Are Thcories of Learning Necessary?” Oryan iza ti un, 1 ( 1 }. 39-60.
fsycfi‹›/‹›gi‹'n/ Sri u•iv. 57, 193—2 1 6. ( 19851. “Mental A ccou n1ing and Consumer Choice,”
A/n rim trip .S‹ ir•nc‹', 4 (3), l 99—2 1 4.
Solomon, R.l. and J.D. Corbil ( 1974), “An Opponc n i Pro-
and Hersh m. s her i» ( i s i ), “An Economic Theory
cess Theory of Moti valion.” N m‹•ri‹‘an Ec io ›mi‹ tr-
of Set f-Cont rod.” Bourn o/ o./ P‹›lii i’‹’o1 ñc‹inrim r, 59
y rat, 68 (6l, i 2—24.
(2). 392—4 10.
Stern. Hawk ins ( 19621, “The Signihcancc of I mpulsc Buying Tim e (I 987). “Fightin g the U rge to Splu rgc,” December
Today,” Journal ‹*{,Wp rkUI i ny, 26 (April). 59-62. 14. Wagner. Alan ( 1959). “The Role of Reinforcement and
Stigler, George J. and Gary S. Becker ( 1977), “De G ustibus Non reinfr›rcement in an ‘Apparent Frustration Effect,’ ”
Non Est Disputandu m,” .4/n ric'iin I.c‘onom/‹' fi‹'i'i‹'w, J‹:iirnal of Ex perunenial P.s t /i‹if‹›gi', 57, 130— 136.
8.9 (2), 76-90. Walster, Elai ne. E. Bcrscheid, and G.W. Walster ( 1973).
Stouffer, S.A., E.A. Sue h man, L.C. De Vin ne y, S.A. Star, "New Directions in Eq uit y R cscarch,” Journal o/ Per-
and s‹inulil; un J Svq ial I•.i ycholi›g , 25, t 5—.1 173.
R.M. Will iams ( 19 49), The . I tnt'i’iran Snldi rr.’ Ad ii ‘t- Wcinberg, Peter and Wol fga ng Gottwold ( 1982). “I mpulsi ve D
ment during ArrtiV Li[e, Vol. 1, Princelon. NJ. Princeton Consumer Bu vi n g as a Rcsuli of E moi ions,” Journul ol o
U ni versity Press. Bii.si n‹ s.s R •.ecu r 'h. 1. 0 (March), 43—57, w
Slrotr, Robert H . ( I 95h1, “M yopia and Inco nsist c ncy in nl
Wells, Will iam D. ( 1987), “Lectures and Dramas,” in C‹›y-
Dynamic U tilitv Maxim izatio n,” R‹ vi ‹ i›’ of’ oa
n ir/v‹ nu dA JLve'l i ’e Re.i-pr›n›’e /‹i 1 di'c•rli s ing, ed. Pat
.$io‹fir.s. 2.3 (2). 1.65 — 180. de
Caf- ferata and A lice Tybo ut, Lex ington, MA: Heath. d
Winston, Gordon C. ( 1980), “Addiction and Backsliding: fro
A Theory of Com pulsivc Consu mption,” .hmmal of m
F.crior›mir’ B‹'huv inr and Organ ical izin. I , 295—324. htt
Yaies, Brian T. and Wa licr M ischcl ( 1979), “Young Chil- ps
d ren's Prcfcrrcd Attentional Strategies for Dela y ing ://
G rati fication.” Joni na I ri/’ Pier.s c'nalit t’ und Swi ft1 P.s ac
v- c-li‹›I‹›y) , 37 t2), 286—3fT(J. ad
Terkel. Si uds ( 19 70). Ha rd “I"iinc.•. .4 n II ral Hi.sinr ’ o[ ter Zajonc, Robert B. (1 968}. “Social Facilitation.” in Grouy e
Gr‹•o/ D‹•pre.i.sion, New York: Pa nthcon. K5'n emit .s, ed. Dorwin Cart wrigh t and Alvin Zander. mi
Thaler, Richard H . ( 1980), “Towurd a Positive Thcor j of New Y ork: lfarpcr & Row. c.
ou
p.
co
m/
jcr
/ar
ticl
e/
17
/4/
49
2/
17
97
24
3
by
gu
es
t
on
09
O