Академический Документы
Профессиональный Документы
Культура Документы
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
American Economic Association is collaborating with JSTOR to digitize, preserve and extend access to The
American Economic Review.
http://www.jstor.org
People are impatient-they like to experi- immediate rewards and to avoid immediate
ence rewards soon and to delay costs until costs in a way that our "long-run selves" do
later. Economists almost always capture im- not appreciate. For example, when presented
patience by assuming that people discount a choice between doing seven hours of an un-
streams of utility over time exponentially. pleasant activity on April 1 versus eight hours
Such preferences are time-consistent: A per- on April 15, if asked on February 1 virtually
son's relative preference for well-being at an everyone would prefer the seven hours on
earlier date over a later date is the same no April 1. But come April 1, given the same
matterwhen she is asked. choice, most of us are apt to put off the work
Casual observation, introspection, and psy- until April 15. We call such tendencies
chological research all suggest that the as- present-biased preferences: When consider-
sumption of time consistency is importantly ing trade-offs between two future moments,
wrong.' It ignores the human tendency to grab present-biasedpreferences give stronger rela-
tive weight to the earlier moment as it gets
closer.2
* O'Donoghue:Departnentof Economics,414 UrisHall,
CornellUniversity,Ithaca,NY 14853;Rabin:Departmentof damental axiom of (rational) human behavior. For some
Economics,549 Evans Hall #3880, Universityof California, recent discussions of empirical evidence of time inconsis-
Berkeley, CA 94720. We thank Steven Blatt, Erik Eyster, tency, see Richard H. Thaler (1991) and Thaler and
and Clara Wang for useful researchassistance,and Steven Loewenstein (1992).
Blatt, Erik Eyster,David Laibson,two anonymousreferees, 2 Many researchershave studiedtime-inconsistentpref-
and seminar participantsat the University of Califomnia- erences. A small set of economists have over the years
Berkeley, NorthwesternUniversity,the Russell Sage Foun- proposed formal, general models of time-inconsistent
dation, Columbia University, Yale University, Harvard preferences. See, for instance, Robert H. Strotz (1956),
University,MIT, the Universityof Wisconsin,CornellUni- E. S. Phelps and Robert A. Pollak (1968), Pollak (1968),
versity, Arizona State University, Santa Clara University, Bezalel Peleg and Menahem I-. Yaari (1973), Yaari
Texas A&M University, and the University of Chicago for ( 1977), and Steven M. Goldman (1979, 1980). Otherre-
helpful comments. For financial support, we thank the searchers have posited a specific functional form, hyper-
National Science Foundation (Grant No. 9709485). bolic discounting, to account for observed tendencies for
O'Donoghuethanksthe Alfred P. Sloan Foundationand the immediate gratification [see Shin-Ho Chung and Richard
Math Centerat NorthwesternUniversity,and Rabin thanks J. Herrnstein (1967), George Ainslie and Herrnstein
the AlfredP. Sloan and Russell Sage Foundations. (1981), Ainslie (1991, 1992), Ainslie and Nick Haslam
' George Loewenstein ( 1992) reviews how the eco- (1992b), and Loewenstein and Drazen Prelec (1992)].
nomics profession evolved from perceiving exponential We have contrived the term "present-biasedpreferences"
discounting as a useful, ad hoc approximation of as a more descriptive term for the underlyinghumanchar-
intertemporal-choicebehavior, to perceiving it as a fun- acteristic that hyperbolic discounting represents.
103
In this paper, we explore the behavioral and In Section I, we further motivate and for-
welfare implications of present-biasedprefer- mally define a simplified form of present-
ences in a simple model where a person must biased preferences [originally proposed by
engage in an activity exactly once duringsome Phelps and Pollak (1968) and later employed
length of time. This simple model encom- by Laibson (1994)] that we study in this pa-
passes an important class of situations, and per: Relative to time-consistent preferences, a
also allows us to lay bare some basic principles person always gives extraweight to well-being
that might apply more generally to formal now over any future moment, but weighs all
models of time-inconsistent preferences. future moments equally. In Section II, we set
Our analysis emphasizes two sets of dis- up our model of a one-time activity. We sup-
tinctions. The first distinction is whether pose that a person must engage in an activity
choices involve immediate costs-where the exactly once during some length of time. Im-
costs of an action are immediate but any re- portantly, at each moment the person can
wards are delayed-or immediate rewards- choose only whether or not to do it now, and
where the benefits of an action are immedi- cannot choose when later she will do it. Within
ate but any costs are delayed. By exploring this scenario, we consider a general class of
these two different settings under the rubric reward and cost schedules for completing the
of present-biased preferences, we unify the activity.
investigation of phenomena (e.g., procras- Section III explores the behavioral impli-
tination and overeating) that have often been cations of present-biased preferences in our
explored separately, but which clearly come model. We present two simple results charac-
from the same underlying propensity for im- terizing how behavior depends on whether re-
mediate gratification.3 wards or costs are immediate, and on whether
The second distinction is whether people people are sophisticatedor naive. The present-
are sophisticated, and foresee that they will bias effect characterizesthe directimplications
have self-control problems in the future, or of present-biased preferences: You procrasti-
are naive, and do not foresee these self- nate-wait when you should do it-if actions
control problems. By explicitly comparing involve immediate costs (writing a paper),
these competing assumptions-each of and preproperate-do it when you should
which has received attention in the econom- wait-if actions involve immediate rewards
ics literature-we hope to delineate which (seeing a movie). Naive people are influenced
predictions come from present-biased pref- solely by the present-bias effect. The sophis-
erences per se, and which come from these tication effect characterizes the direct impli-
assumptions about foresight.4 cations of sophistication versus naivete: A
sophisticated person does the activity sooner
3 Throughout this paper, our emphasis is impulsive
than does a naive person with the same pref-
choice driven by a tendency to overweight rewards and erences, irrespective of whether rewards or
costs that are in close temporal proximity. But there are costs are immediate. Intuitively, a sophisti-
clearly other aspects of impulsive choice as well: People cated person is correctly pessimistic about her
also tend to overweight rewards and costs that are in close future behavior-a naive person believes she
spatial proximity, and more generally are attentive to re-
wards and costs that are salient (see Loewenstein, 1996). will behave herself in the future while a so-
4 Strotz (1956) and Pollak (1968), two of the seminal phisticated person knows she may not. As a
papers on time-inconsistent preferences, carefully lay out result, waiting always seems less attractivefor
these two assumptions, but do not much consider the im- a sophisticated person. Although the direction
plications of one versus the other. More recentpapershave
assumed either one or the other, without attemptingto jus-
is the same, the sophistication effect has very
tify the choice on behavioral grounds. For instance, differentconnotationsfor immediatecosts ver-
George A. Akerlof (1991) assumes naive beliefs, while sus immediate rewards. When costs are im-
David Laibson (1994, 1995, 1997) and Carolyn Fischer mediate, sophistication mitigates the tendency
( 1997) assume sophisticated beliefs. Each paper states its to procrastinate.(And in fact, the sophistica-
assumption about beliefs used [and Akerlof ( 1991 ) posits
that his main welfare finding depends on his assumption tion effect can outweigh the present-biaseffect
of naive beliefs], but conspicuously does not argue why so that a sophisticated person may perform an
its assumption is correct. onerous activity before she would if she had
no self-control problem.) When rewards are fare loss from a small present bias is small if
immediate, on the other hand, sophistication you are naive.
exacerbates the tendency to preproperate. Researchers looking for empirical proof of
In Section IV, we tum to the welfare re- time-inconsistent preferences often explore
sults.5 Again, the two distinctions-immedi- the use of self-limiting "commitment de-
ate costs vs. immediate rewards and vices" (e.g., Christmasclubs, fat farms), be-
sophistication vs. naivete-are crucial. When cause such devices represent"smoking guns"
costs are immediate, a person is always better that cannot be explained by any time-
off with sophisticated beliefs than with naive consistent preferences. We show in Section V
beliefs. Naivete can lead you to repeatedly that even within our simple setting, certainbe-
procrastinatean unpleasant activity under the haviors induced by present-biasedpreferences
incorrect belief that you will do it tomorrow, are inconsistent with any time-consistentpref-
while sophistication means you know exactly erences. Hence, we illustrate that smoking
how costly delay would be. In fact, even with guns need not involve external commitment
an arbitrarilysmall bias for the present, for devices. Furthermore, while previous litera-
immediate costs naive people can experience ture has focused on smoking guns for sophis-
severe welfare losses, while the welfare loss ticated people, we show that smoking guns
from a small present bias is small if you are exist for naive people as well.
sophisticated. When rewards are immediate, Although many of the specific results de-
however, a person can be better off with naive scribed above are special to our one-activity
beliefs. In this case, people with present- model, these results illustrate some more gen-
biased preferencestend to do the activity when eral intuitions. To begin the process of gen-
they should wait. Naivete helps motivate you eralizing our model, in Section VI we present
to wait because you overestimate the benefits an extension where, rather than being per-
of waiting. Sophistication makes you (prop- formed exactly once, the activity must be per-
erly) skeptical of future behavior, so you are formed more than once during some length of
more tempted to grab today's immediate re- time. In Section VII, we discuss more broadly
ward. This can lead to "unwinding" similar (and less formally) what our model suggests
to that in the finitely repeated prisoner's di- about general implications of self-control
lemma: In the end, you will give in to temp- problems, and describe how some of these im-
tation and grab a rewardtoo soon; because you plications might play out in specific economic
realize this, near the end you will cave in a contexts, such as saving and addiction. We
little sooner than if you thought you would then conclude with a discussion of some les-
resist temptationin the end; realizing this, you sons to take away from our analysis, both for
will cave in a little sooner, etc. As a result, for why it is important that economists start to
immediate rewards it is sophisticated people study self-control problems, and for how we
who can experience severe welfare losses with should go about doing so.
an arbitrarilysmall present bias, while the wel-
L. Present-BiasedPreferences
simple model employed by economists is ex- Definition 1: (3, 6) -preferences are prefer-
ponential discounting: For all t, Ut (ut, u 1I, ences that can be representedby:
UT) T 6 Tu, where 8 E (0, 1] is a
'discount factor." For all t, Ut(ut u,g+
u, 1e U7)
Exponential discounting parsimoniously
capturesthe fact that people are impatient.Yet
exponential discounting is more than an in- =6'u, +8 67-uT
nocuous simplification of a more general class r- t+ I
clinics, Christmasclubs, or fat farms, provides but incurs cost ct, and makes no further
evidence of sophistication.9Only sophisticated choices. If she waits, she then will face the
people would want to commit themselves to same choice in period t + 1. Importantly, if
smaller choice sets: If you were naive, you the person waits she cannot commit in period
would never worry that your tomorrow self t to when later she will do it. If the person waits
might choose an option you do not like today. until period T, she must do it then.
Despite the existence of some sophistication, The rewardschedule v and the cost schedule
however, it does appear that people underes- c representrewards and costs as a function of
timate the degree to which their futurebehav- when the person does the activity. However,
ior will not match their current preferences the person does not necessarily receive the re-
over futurebehavior. For example, people may wards and costs immediately upon completion
repeatedly not have the "will power" to forgo of the activity. Indeed, we differentiate cases
tempting foods or to quit smoking, while pre- precisely by when rewards and costs are ex-
dicting that tomorrow they will have this will perienced. Some activities, such as writing a
power. We think there are elements of both paper or mowing the lawn, are unpleasant to
sophistication and naivete in the way people perform, but create future benefits. We refer
anticipate their own futurepreferences. In any to activities where the cost is incurred imme-
event, our goal is to clarify the logic of each, diately while the rewardis delayed as activities
and in the process we delineate which pre- having immediate costs. Other activities, such
dictions come purely from present-biased as seeing a movie or taking a vacation, are
preferences, and which come from the "so- pleasurable to perform, but may create future
phistication effects" of people being aware of costs. We refer to activities where the reward
their own time inconsistency. is received immediately while the cost is
delayed as activities having immediate
IL. DoingIt Once rewards.'?
We analyze these two cases using the (Q, 8)-
Suppose there is an activity that a person preferences outlined in Section 1. For simplic-
must performexactly once, and thereare T < oo ity, we assume 6 = 1; i.e., we assume thatthere
periods in which she can do it. Let v (v1, is no "long-term" discounting." Given 8 = 1,
V2, ...VT) be the reward schedule, and let without loss of generality we can interpretde-
C (C1, C2... CT) be the cost schedule, where layed rewardsor costs as being experienced in
vt Oand c 2 0 for each E t {1, 2, ... , T}. period T + 1. We can then describe a person's
In each period t - T - 1, the person must intertemporal utility from the perspective of
choose either to do it or to wait. If she does period t of completing the activity in period
the activity in period t, she receives rewardvt J 2 t, which we denote by U't(r).12
mists. Akerlof (1991) uses a variant of the naivete " We occasionally make reference to a thirdcase where
assumption. both rewards and costs are immediate. The fourth case-
'The very term "self-control" implies that people are neither rewardsnor costs are immediate-is not of inter-
aware that it may be prudentto control their future selves. est because it is equivalent to the case of time consistency,
For analyses of self-control in people, see Ainslie (1974, which we study.
1975, 1987, 1992), Thomas C. Schelling (1978, 1984, " The results are easily generalized to 6 < 1. Suppose
1992), Thaler (1980), Thaler and Hersh M. Shefrin the "true" reward schedule is ir (7r1, 7r2, . rT), the
( 1981), David C. Funderand Jack Block (1989), Stephen "true" cost schedule is 4 (41 <2,
4.. T), and 6 < 1.
J. Hoch and Loewenstein (1991), Ainslie and Haslam If, for instance, costs are immediate and rewards are re-
(1992a), Jacob Glazer and Andrew Weiss (1992), ceived in period T + 1, then if we let vt - 6T+ It and
Shefrin and Thaler (1992), Klaus Wertenbroch (1993), 64-P&,for each t, doing the analysis with v, c, and no
and Laibson (1994, 1995, 1997). Ainslie (1974) explores discounting is identical to doing the analysis with ir, 4
similar issues with pigeons. As many have emphasized, and 6.
especially Ainslie (1992) and Bill Watterson (1993 pp. 12This formulationnormalizes the instantaneousutility
83-88), a sort of intrapersonal "bargaining" can arise from not completing the activity to be zero. For instance,
because of the basic disagreementswe have with ourselves when costs are immediate and rewards are received in pe-
about when we should do something. riod T + 1, we are assuming that if the person does the
1. Immediate Costs. -If a person com- To describe behavior given our assump-
pletes the activity in period r, then her inter- tions, we define a "solution concept": A
temporal utility in period t G T is perception-perfectstrategy is a strategythatin
all periods (even those after the activity is per-
Ut(-r) 3
F3v, - c if r t formed) a person chooses the optimal action
given her currentpreferences and her percep-
l ,T- fcT if r > t. tions of future behavior. Rather than give a
general formal definition, we simply define a
2. ImmediateRewards.-If a person com- perception-perfect strategy for each of the
pletes the activity in period T, then her inter- three types of agents that we consider. Defi-
temporalutility in period t G T is nition 2 describes a perception-perfectstrategy
for TCs. Reflecting the fact that TCs do not
VT - 6cT if r = t have a self-control problem, Definition 2 says
that in any period, TCs will complete the ac-
,v-
#V 6cT if 'r > t. tivity if and only if it is the optimal period of
those remaininggiven her currentpreferences.
We will focus in this environment on three
types of agents. We refer to people with stan- Definition 2: A perception-perfect strategy
dard exponential, time-consistent preferences for TCs is a strategySIC (s ,s, .6x, s72) that
.
(i.e., 8 = 1) as TCs. We then focus on two satisfies for all t < T Stf = Y if and only if
types of people with present-biased prefer- Ut(t) 2 Ut(r) for all r > t.
ences (i.e., ,B < 1), representing the two ex-
tremes discussed in Section I. We call people Naifs have present-biased preferences
with sophisticated perceptions sophisticates, (since 13< I), but naifs believe that they are
and people with naive perceptions nazfs. So- time-consistent. As a result, the decision pro-
phisticates and naifs have identicalpreferences cess for naifs is identical to that for TCs (al-
(throughout we assume they have the same though naifs have different preferences).
/3), and therefore differ only in their percep- Definition 3 says that in any period, naifs will
tions of future preferences. complete the activity if and only if it is the
A person's behavior can be fully described optimal period of those remaining given her
by a strategy = (s s2, ..2, ST), where s, E currentpreferences.
{Y, N} specifies for period t E 1, 2, ..., T}
whether or not to do the activity in period t Definition 3: A perception-perfect strategy
given she has not yet done it. The strategy s for naifs is a strategy s" =(s7I, sS,
n., SI') that
specifies doing it in period t if st = Y, and satisfies for all t < T sn = Y if and only if
waiting if s, = N. In addition to specifying Ut(t) ? Ut(r) for all T > t.
when the person will actually complete the ac-
tivity, a strategyalso specifies what the person Although naifs and TCs have essentially the
"would" do in periods after she has already same decision process, it is importantto real-
done it; e.g., if s, = Y, we still specify s,' for ize that naifs have incorrectperceptions about
all t' > t. This featurewill prove useful in our future behavior, and therefore may plan to be-
analysis. Since the person must do it in period have one way but in fact behave differently.
T if she has not yet done it, without loss of With (#, b)-preferences, these incorrect per-
generality we require ST = Y- ceptions take a convenient form: At all times,
naifs believe that if they wait they will behave
like TCs in the future.
Sophisticates also have present-biasedpref-
activity in period r, the instantaneous utilities are u1- erences and a self-control problem. But unlike
-C., u7+I = v1-, andu, = for all t Z {T, T + 1}. This naifs, sophisticates know they will have self-
assumption is purely for convenience. In particular,for
any i we would get identical results if we normalize the
control problems in the future, and therefore
utility from not doing the activity to be ii, or if we nor- correctly predict future behavior. Definition 4
malize the utility from completing the activity to be i. says that in period t, sophisticates calculate
when their future selves will complete the ac- lowing scenario: Suppose you usually go to
tivity if they wait now, and then do the activity the movies on Saturdays, and the schedule at
now if and only if given their currentprefer- the local cinema consists of a mediocre movie
ences doing it now is preferredto waiting for this week, a good movie next week, a great
their future selves to do it. movie in two weeks, and (best of all) a Johnny
Depp movie in three weeks. Now suppose you
Definition 4: A perception-perfect strategy must complete a report for work within four
for sophisticates is a strategy ss _ (sS, s2- weeks, and to do so you must skip the movie
Ss) that satisfies for all t < T ss = Y if and on one of the next four Saturdays. When do
only if U'(t) 2 U'(T') where T' min,,, you complete the report?
{TIs -- Y}. The activity you must do exactly once is
writing the report. The reward from doing the
Note that in Definitions 2, 3, and 4, we have report is received at work in the future. We
assumed that people do it when indifferent, will assume the rewardis independentof when
which implies that there is a unique you complete the report, and denote it by v.
perception-perfect strategy for each type. In The cost of doing the report on a given Sat-
addition, this assumption implies that a urday--not seeing the movie shown thatday-
perception-perfect strategy must be a pure is experienced immediately. Letting valua-
strategy. For generic values of v, c, and /, tions of the mediocre, good, great, and Depp
nobody will ever be indifferent, so these as- movies be 3, 5, 8, and 13, we formalize this
sumptions are irrelevant.In nongeneric games, situation in the following example, where we
more general definitions could lead to present both the parameters of the example
additional equilibria. For sophisticates, a and the perception-perfect strategy for each
perception-perfectstrategy is the identical so- type of agent.
lution concept to that used by Strotz (1956),
Pollak (1968), Laibson (1994, 1995, 1997), Example 1: Suppose costs are immediate,
and others. For naifs, it is essentially the same T = 4, and ,6 = 1/2for naifs and sophisticates.
solution concept as those used by Pollak Let v = (v, v, v, v ) and c = (3, 5, 8, 13).
(1968) and Akerlof (1991) .
It will be useful in the analysis of this model s' (Y, Y, Y, Y), so TCs do the report in
to have notation for when a person will actu-
ally complete the activity (i.e., the outcome): period rtc = 1.
Given the perception-perfectstrategies s"c,sS
and Sn, we let Tr,, rs, and Tn be the periods in sn (N, N, N, Y), so naifs do the report in
which each of the three types of agents do the
activity. That is, given a E { tc, s, n }, Tra period T-n= 4.
min,{tis' = Y}.
s= (N, Y, N, Y), so sophisticates do the
III. Behavior
report in period rs = 2.
In this section, we compare the behavior of
TCs, naifs, and sophisticates who have iden- TCs do the reporton the first Saturday,skip-
tical long-run preferences. Comparingnaifs or ping the mediocre movie. TCs always do the
sophisticates to TCs reflects how people with activity in the period t that maximizes vt - ct.
present-biased preferences behave relative to Since Example 1 has a stationary reward
how they would like to behave from a long- schedule, TCs do the reportin the period with
run perspective; and comparing sophisticates the minimum cost.
to naifs reflects the implications of sophisti- Naifs procrastinateuntil the last Saturday,
cation about self-control problems. forcing themselves to skip the Depp movie. On
We begin by analyzing in some detail a pair the first Saturday, naifs give in to their self-
of related examples to illustrate the intuitions control problem and see the mediocre movie
behind many of the results. Consider the fol- because they believe they will skip the good
movie in week 2 and still be able to see the for a movie. The scledule at the local cinema
great movie and the Depp movie. The period-I is the same as for the above example-a me-
naif prefers incurninga cost of 5 next week as diocre movie this week, a good movie next
opposed to a cost of 3 now. However, when week, a great movie in two weeks, and (best
the second Saturday arrives, naifs again give of all) a Johnny Depp movie in three weeks.
in to their self-control problem and see the Which movie do you see?
good movie, now believing they will skip the Now, the activity you must do exactly once
great movie in week 3 and still get to see the is going to a movie, and the reward, seeing the
Depp movie. Finally, when the third Saturday movie, is experienced immediately.'3 Using
aTirves,naifs have self-control problems for a the same payoffs for seeing a movie as in Ex-
third time and see the great movie, forcing ample 1, we have the following formalization.
themselves to miss the Depp movie. This ex-
ample demonstratesa typical problemfor naifs Example 2: Suppose rewards are immediate,
when costs are immediate: They incorrectly T = 4, and 3 B/2for naifs and sophisticates.
predict that they will not procrastinatein the Let v - (3, 5, 8, 13) and c (0, 0, 0, 0).
future, and consequently underestimate the
cost of procrastinatingnow. Stc (N, N, N, Y), so TCs see the movie in
Sophisticates procrastinate one week, but
they do the report on the second Saturday, period -tc - 4.
skipping the good movie and enabling them-
selves to see the great movie and the Depp sn (N, N, Y, Y), so naifs see the movie in
movie. The period-I sophisticate correctly
predicts that he would have self-control prob- period -, = 3.
lems on the third Saturday and see the great
movie. However, the period-i sophisticate = (Y, Y, Y, Y), so sophisticates see the
also correctly predicts that knowing about
period-3 self-control problems will induce him movie in period -r = 1.
to do the report on the second Saturday.
Hence, the period-i sophisticate can safely TCs wait and see the Depp movie since it
procrastinateand see the mediocre movie. Ex- yields the highest reward.Naifs see merely the
ample 1 illustratestypical behavior for sophis- great movie. On the first two Saturdays,naifs
ticates when costs are immediate. Although skip the mediocre and good movies incorrectly
sophisticates have a tendency to procrastinate believing they will wait to see the Depp movie.
(they do not write the reportright away, which However, on the third Saturday, they give in
their long-run selves prefer), perfect foresight to self-control problems and see the great
can mitigate this problem because sophisti- movie. For activities with immediate rewards,
cates will do it now when they (correctly) the self-control problem leads naifs to do the
foresee costly procrastinationin the future. activity too soon.
Example 1 illustrates an intuition expressed Sophisticates have even worse self-control
by Strotz (1956) and Akerlof (1991) that so- problems in this situation. They see merely the
phistication is "good" because it helps over- mediocre movie because of an unwinding sim-
come self-control problems. As in Akerlof's ilar to that in the finitely repeated prisoner's
(1991) procrastinationexample, naifs repeat- dilemma. The period-2 sophisticate would
edly put off an activity because they believe choose to see the good movie because he cor-
they will do it tomorrow. Akerlof intuits that rectly predicts that he would give in to self-
sophistication could overcome this problem, control problems on the third Saturday, and
and Example 1 demonstratesthis intuition. see merely the great movie rather than the
However, this intuition may not hold when
rewardsare immediate. Consider a similarsce-
nario: Suppose you have a coupon to see one '" That seeing a movie is a "cost" in Example 1 and a
movie over the next four Saturdays, and your "'reward"in Example 2 reflects that the rewardsand costs
allowance is such thatyou cannot affordto pay are defined with respect to the activity being done once.
Depp movie. The period-I sophisticate cor- immediate costs naifs always procrastinate,
rectly predicts this reasoning and behavior by and for immediate rewards naifs always
his period-2 self. Hence, the period-I sophis- preproperate.17
ticate realizes that he will see merely the good
movie if he waits, so he concludes he might as PROPOSITION 1: (1) If costs are immedi-
well see the mediocre movie now. This ex- ate, then i-, ?r,2
ample demonstrates a typical problem for so- (2) If rewards are immediate,then Ttt. Tn C
Because sophisticates are influenced by the self-control problem. Even so, situations
the sophistication effect in addition to the like that in Example 3 are not particularly
present-bias effect, the qualitative behavior pathological, and "preemptive overcontrol"
of sophisticates relative to TCs is compli- is likely to arise in real-world environments
cated. In particular, it can be that sophisti- (especially when choices are discrete). We
cates do not even exhibit the basic highlight this result to emphasize the impor-
present-bias intuition. Consider the follow- tance of sophistication effects. If you assume
ing scenario: Suppose you must write a pa- present-biased preferences and sophistication
per this weekend, on Friday night, Saturday, (as economists are prone to do), you must be
or Sunday. You know the paper will be bet- careful to ask whether results are driven by
ter if written on either Saturday or Sunday present-biased preferences per se, or by
(when you have an entire day). However, it present-biasedpreferences in conjunctionwith
is a mid-November weekend with plenty of sophistication effects.
sports on TV-pro basketball on Friday
night, college football on Saturday, and pro
IV. Welfare
football on Sunday. You prefer watching pro
football to college football, and prefer col-
lege football to pro basketball. Which sports Our emphasis in the previous section on
event do you miss to write the paper? We qualitative behavioral comparisons among
can represent this scenario with the follow- the three types of people masks what we feel
ing example, where the activity to be done may be a more important question about
once is writing the paper and the costs cor- present-biased preferences: When does the
respond to the attractiveness of the sports taste for immediate gratification severely
event missed. hurt a person? In this section, we examine
the welfare implications of present-biased
Example 3: Suppose costs are immediate, preferences with an eye towards this ques-
T = 3, and /3 = 1/2 for naifs and sophisticates. tion. We show that even a small bias for the
Let v = {12, 18, 181 andc = {3, 8, 13}. present can lead a person to suffer severe
ThenTs = 1 andTt- = 2 (and Tn = 3) welfare losses, and characterize conditions
when this can happen.
TCs write the paperon Saturdaybecause the Welfare comparisons for people with time-
marginal benefit of a better paper outweighs inconsistent preferences are in principle prob-
the marginalcost of giving up college football lematic; the very premise of the model is that
for pro basketball. Since the example involves a person's preferences at different times dis-
immediate costs, the present-bias effect sug- agree, so that a change in behavior may make
gests that sophisticates should procrastinate. some selves better off while making other
However, the sophistication effect leads so- selves worse off. The savings literature(e.g.,
phisticates to write the paper on Friday night, Goldman, 1979, 1980; Laibson, 1994) often
before TCs. On Friday, sophisticates correctly addresses this issue by defining a Pareto-
predict that they will end up writing the paper efficiency criterion, asking when all period
on Sunday if they do not do it now. Hence, selves (weakly) prefer one strategyto another.
although sophisticates would prefer to write If a strategy is Pareto superiorto another,then
the paper on Saturday,they do it on Friday to it is clearly better. However, we feel this cri-
prevent themselves from procrastinatinguntil terion is too strong: When applied to intertem-
Sunday. poral choice, the Pareto criterionoften refuses
In Example 3, sophisticates behave exactly to rank two strategies even when one is much
opposite from what present-biasedpreferences preferredby virtually all period selves, while
would suggest, a result we will see again in the other is preferredby only one period self.
Sections VI and VII. Of course, this is not al- Since present-biased preferences are often
ways the case. Indeed, when rewards are im- meant to capture self-control problems, where
mediate, sophisticates always preproperate people pursue immediate gratification on a
because the sophistication effect exacerbates day-to-day basis, we feel the naturalperspec-
tive in most situations is the "long-run perspective (i.e., it also has an even larger de-
perspective". (See Schelling [1984] for a layed cost), then more severe preproperation
thoughtful discussion of some of these by sophisticates may in fact mean that sophis-
issues.) 18 ticates avoid this "temptation trap" while
To formalize the long-run perspective, we naifs do not. Hence, for immediate rewardswe
suppose there is a (fictitious) period 0 where cannot say in general whether sophisticates or
the person has no decision to make and naifs are better off.
weights all futureperiods equally. We can then Rather than simple comparisons between
denote a person's long-run utility from doing sophisticates and naifs, however, our main fo-
it in period r by U?(Tr) v) - cT. Ourwelfare cus for welfare analysis is the question of
analysis throughout this section will involve when a small bias for the present (i.e., j3 close
comparisons of long-run utilities. Even so, to 1) can cause severe welfare losses. Since
most of our welfare comparisons can be sophisticates, naifs, and TCs have identical
roughly conceived of as "Pareto compari- long-run utility, we can measure the welfare
sons,"' and we will note Pareto-efficiency loss from self-control problems by the devia-
"'analogues" for our two main welfare results tion from TC long-run utility [i.e., U'(-r,,) -
at the end of this section. U0J(T5)and U0(Tr,) -U(r)].
We begin with some brief qualitative com- We first note that if rewards and costs can
parisons of sophisticates and naifs. The lan- be arbitrarilylarge, then a person with present-
guage in Section III implied that sophistication biased preferences can suffer arbitrarilysevere
is good when costs are immediate because it welfare losses even from one-shot decisions.
mitigates the tendency to procrastinate. In- Suppose rewards are immediate, for instance,
deed, it is straightforwardto show that when in which case a person with present-biased
costs are immediate, sophisticates always do preferences is willing to grab a reward today
at least as well as naifs [i.e., U (-r,) ? for a delayed cost that is largerthan the reward
U0(rn )I. Intuitively, since sophisticates never (by factor 1/ /3). Even if /3 is very close to one,
procrastinatein a period where naifs do it, the this decision can create an arbitrarily large
only way their utilities can differ is when so- welfare loss if the reward and cost are large
phisticates preempt costly procrastination. enough.
When sophisticates choose to preempt costly We feel the more interesting case is when
procrastination, they do so despite their ex- there is an upper bound on how large rewards
aggerated aversion to incurring immediate and costs can be. In this case, the welfare loss
costs, so this decision must also be preferred from any individual bad decision will become
by the long-run self. very small as the self-control problem be-
When rewards are immediate, on the other comes small. But even if the welfare loss from
hand, the discussion in Section III implied that any individual decision is small, severe wel-
sophisticationis bad because it exacerbatesthe fare losses can still arise when self-control
tendency to preproperate.More severe prepro- problems are compounded. To demonstrate
peration will often lead to lower long-run util- this result, we suppose the upperbound on re-
ity (as in Example 2), but this is not wards and costs is X. Then the welfare loss for
necessarily the case. In particular,if there is a both soplhisticatesand naifs cannot be larger
future period that is very tempting (i.e., it has than 2X.
a large reward) but very bad from a long-run Co-nsiderthe case of immediate costs, where
the self-control problem leads you to procras-
tinate. As in Example 1, naifs can compound
self-control problems by making repeated de-
8 Indeed, Akerlof (1991) frames his discussion of pro- cisions to procrastinate, each time believing
crastinationin a way that emphasizes that a person's true they will do it next period. With each decision
preferences are her long-run preferences. Procrastination to procrastinate, they incur a small welfare
occurs in his model because costs incurredtoday are "sa-
lient" -a person experiences a cognitive illusion where loss, but the total welfare loss is the sum of
costs incurred today loom larger than they are according these increments. No matterhow small the in-
to her true preferences. dividual welfare losses, naifs can suffer severe
welfare losses if they procrastinate enough terion to use when examining the welfare im-
times. Sophisticates, in contrast,know exactly plications of present-biasedpreferences.Using
when they will do it if they wait, so delaying this criterion, Propositions 3 and 4 formalize
from period wr to period -r is a single decision when a small bias for the present can be very
to procrastinate.Hence, for sophisticates small costly from a long-run perspective.'9Even so,
self-control problems cannot cause severe we note that there is also a less strong formal-
welfare losses. The following proposition for- ization using Pareto comparisons: If costs are
malizes these intuitions. immediate, sophisticates always choose a
Pareto-optimal strategy while naifs may not;
PROPOSITION 3: Suppose costs are imme- and if rewards are immediate, naifs always
diate, and consider all v and c such that v, s X choose a Pareto-optimal strategy while so-
and c, ? Xfor all t: phisticates may not.
(1) limel1(sup(v,C)[U0(r,) -Uo(r)]) =
0, and V. SmokingGuns
(2) For any 3 < 1,
U(T,) - U(Tr,)] -2X.
sup(V,C)[ Many researchers studying time-
inconsistent preferences have searched for
When rewardsare immediate, however, and
the self-control problem leads you to prepro-
perate, we get the exact opposite result. For
" We feel that these limit results qualitatively capture
immediate rewards, naifs always believe that
if they wait they will do it when TCs do it, so very real differences in when moderately impatient so-
phisticates and naifs can suffer severe welfare losses, but
doing it in period Tn as opposed to waiting there are reasons to be cautious in interpretingthem too
until period rt, is a single decision to prepro- literally. For instance, since "unwinding" drives severe
perate for naifs. Hence, for naifs small self- preproperationfor sophisticates, it seems natural to ask
control problems cannot cause severe welfare whether a small amount of uncertaintycould reverse this
losses. But sophisticates can compound self- tendency, much as David M. Kreps et al. (1982) showed
that a small amount of uncertainty can lead to extensive
control problems because of an unwinding: In cooperation in the finitely repeated prisoner's dilemma.
the end, sophisticates will preproperate;be- We suspect that there is something to this story, but the
cause they realize this, near the end they will analogy is problematicon two fronts. First, althoughplay-
preproperate;realizing this they preproperate ers may cooperate for most of a very long horizon, there
is still a long durationat the end of the repeatedprisoner's
a little sooner, etc. For each step of this un- dilemma where players are unlikely to cooperate. Such an
winding, the welfare loss may be small, but Iendgame" could still create significant welfare losses.
the total welfare loss is the sum of multiple Second, in the Kreps et al. result a player's currentbehav-
steps. As with naifs and immediate costs, no ior will signal something about her future behavior to
matterhow small the individual welfare losses, otherplayers. Since each "player" in our game plays only
once, the comparable signal is that a person in period t
sophisticates can suffer severe welfare losses infers somethingaboutthe propensityof her period-(t + I )
if the unraveling occurs over enough periods. self to wait from the fact that her period-(t - 1) self
These intuitions are formalized in Proposition waited, which requiresthat the period-t self does not know
4, /3. While we believe that such self-inference and self-
signaling go on, there are many issues to be worked out
to understandthe strategic logic and psychological reality
PROPOSITION 4: Suppose rewards are im- of such phenomena.
mediate, and consider all v and c such that A comparableworry about our extreme results for naifs
vt ? X and ct c Xfor all t: is that they will eventually learn that they have a tendency
to procrastinate.Again, we think there is something to this
(1) lim'6l (sup(V,C)[U0(rfC) -U?(r)]) intuition, but we suspect the issue is complicated. The is-
0, and sue of self-inference again arises. Further,people seem to
(2) For any/ < 1, have a powerful ability not to apply general lessons they
sup(v,c)[U?(QrtC) - Uo( r,) =-2X . understandwell to specific situations. For instance, we are
all familiar with the sensation of being simultaneously
aware that we tend to be overoptimistic in completing pro-
As discussed at the beginning of this sec- jects, but still being overoptimistic regarding our current
tion, we feel that examining welfare losses in project. (See Kahnemanand Dan Lovallo [1993] for ev-
terns of long-run utility is the appropriatecri- ideniceon related issues.)
empirical proof that people have such prefer- A person's behavior is independent of irrele-
ences. Efforts to indirectly prove time incon- vant alternatives if whenever she chooses pe-
sistency have focused on the use of external riod T' * t when facing v and c she also
"commitment devices" that limit future chooses T' when facing v -'and c -t.
choice sets, because the use of such devices
provides smoking guns that prove time con- A time-consistent person will never vio-
sistency wrong. In this section, we show that late dominance nor independence of irrele-
smoking guns exist in our simple one-activity vant alternatives. These results hold for
model, where no externalcommitmentdevices any time-consistent preterences, including
are available. time-consistent preferences that discount
There are two properties that a person differently from period to period, and even
with time-consistent preferences will never time-consistent preferences that are not ad-
violate. The first is "dominance": For inter- ditively separable. Proposition 5 establishes
temporal choice, one strategy dominates an- that these results do not hold for people with
other if it yields in every period an present-biased preferences.
instantaneous utility at least as large as the
instantaneous utility from the other strategy, PROPOSITION 5: For any /3 and 8 such that
and strictly larger for some periods. In our o< 6 c 1 and 0 < f3 < 1, and for both so-
model, one strategy is dominated by another phistication and naivete:
if and only if the first strategy implies doing (1) There exists (v, c) and assumptions
it at a cost with no reward while the second about immediacy such that a person with
strategy implies doing it for a reward with (p6, 8)-preferences will violate dominance,
no cost.20 and
(2) There exists (v, c) and assumptions
Definition 5: A person obeys dominance about immediacy such that a person with
if whenever there exists some period T with (/51,b)-preferences will violate independence
VT > 0 and CT = 0 the person does not do it in of irrelevant alternatives.
any period T' with CT/> 0 and VT' = 0.
To give some intuition for these results, we
The second property that a person with describe examples where each type violates
time-consistent behavior will never violate is dominance. The intuition for why each type
independence of irrelevant alternatives- violates independence of irrelevant alterna-
eliminating an option from the choice tives is related. Sophisticates violate domi-
set that is not chosen should not change nance when they choose a dominated early
the person's choice from the remaining time to do an activity because they (correctly)
options. worry that their future selves will not choose
the dominating later time. For example, sup-
Definition 6: For any v (vI, V2, .-., VT) pose rewards are immediate, T = 3, v = (0, 5,
and c (c,, C2, ..., CT), define 1) and c = (1, 8, 0). Doing it in period 1 is
clearly dominatedby doing it in period 3. Even
v-t (V1. . Vt - 1, Vt + 1, .. . VT) and so, a sophisticate with fi = 1/2 will complete
the activity in period 1. She does so not be-
Ct - (cl. C*- Ct-1C Ct+ *1- , CT). cause it is her most preferredperiod, but rather
to avoid doing it in period 2. In period 1, the
person prefers period 3 to period 1. Unfortu-
nately, the period-2 self gets to choose be-
tween periods 2 and 3, and she will choose
20 E.g., consider a three-periodexample where v = (1,
period 2.
x, 0) and c = (0, y, 1). Then if costs are immediate, doing Naifs can violate dominance because of in-
it in period 1 yields the stream of instantaneousutilities
(0, 0, 0, 1) while doing it in period 3 yields the streamof correct perceptions about future behavior. For
instantaneous utilities (-1, 0, 0, 0). Clearly the former example, suppose costs are immediate, T = 3,
dominates the latter. V = (1, 8, 0) and c = (0, 5, 1). Doing it in
period 3 is dominated by doing it in period 1, in Section II, but now the person must do the
and yet a naif with /1 1/= will choose period activity exactly M 2 1 times, and she can do it
3. Even though in period 1 she prefers period at most once in any given period. We let r i(M)
1 to period 3, she waits in period 1 incorrectly denote the period in which a person completes
believing she will do it in period 2. Unfortu- the activity for the ith time, and define 0(M)
nately, in period 2 she prefers waiting until { Tr'(M), T2(M), ..., TM(M) }. For each period
period 3.21 -r in which the person does it, she receives re-
Proposition 5 has important implications ward vT and incurs cost C, and these can be
for the literature on smoking guns. First, experienced immediately or with some delay.
Proposition 5 implies that smoking guns need Using the interpretationsof immediatecosts and
not involve the use of external commitment immediaterewardsfrom Section II, preferences
devices. Even simple behaviors can some- take the following form.
times represent smoking guns. Furthermore,
the literatureon external commitment devices 1. Immediate Costs. -Given 0(M), the
provides smoking guns for sophisticates but set of periods in which she does it, a person's
not for naifs, since naifs would not pay to intertemporal utility in period t is given by
limit future choice sets. Proposition 5 implies equation (1) below.
that smoking gulls exist for naifs as well. Fi-
nally, the intuitions above (and in the proof) 2. Immediate Rewards. -Given 0( M),
suggest ways to design experiments attempt- the set of periods in which she does it,
ing to find smoking guns, as well as the types a person's intertemporal utility in period
of real world situations without external com- t is given by equation (2) below.
mitment devices where smoking guns might
be found. Given these preferences, we can define
perception-perfect strategiesanalogouslyto Def-
initions 2, 3, and 4. We omit the formal
VI. Multi-Tasking definitions here. Let 0a(M) - {r'(M),
am(M)I be the set of periods that an agent of
We now begin to explore how our results type a E t tc, s, n I completes the activity ac-
might carryover to more generalsettings. Con- cordingto herperception-perfect strategy.We be-
sider a simple extension of our model where gin by showingthatthe behaviorof TCs andnaifs
the activity must be performedmore thanonce. in the multiactivity model is "normal" and
The basic structureof the model is exactly as intuitive.
(I O), +
0ifET (M)
lT 0(M) E(M)t0(M) 7(M )(M) if
tp ( EVT C7).. if t (t ( M) .
TcC (M) Tc OH(M)
(2) Ut(0(M)) 1if)?/( V TCMc)it(M
c
0(-)Vt +p Y,E T
CT) ift O(M)
| ~~~~T
C- (o(M) T-e ((M)
(2) Ut(0(M) )-
KTE VTM CT) if t tO(M).
TC 6M T E- 0-(M)
PROPOSMIION6: (1) For all cases and for alogue to Part I of Proposition 6 does not hold
any v and c, for each M E {1, 2, ..., T 1: for sophisticates. Sophisticates are always
0tc(M) C (4t(M + 1) and 0,(M) C looking for ways to influence their future be-
O(M + 1); and havior, and for M > 1 waiting can be a sort of
(2) If costs are immediate, thenfor all i E "commitment device" to influence future be-
{1, 2, ..., M}, T'T(M) 2 rt(M), and if re- havior. If there is only one activity, there is no
wards are immediate, then for all i C {1, 2, way to commit future selves not to preproper-
.., M 9 'Tti(M) -< rit,(M) . ate. In Example 4, when M 1 the period-I
sophisticate does the activity because he (cor-
Part 1 of Proposition 6 addresses how be- rectly) predicts that he will just do it in period
havior depends on M: If TCs or naifs must do 2 if he waits. If there is a second activity, how-
the activity an extra time, they do it in all pe- ever, a commitment device becomes available:
nods they used to do it, and some additional Waiting now prevents you from doing the ac-
period. If in any period they have k activities tivity for the second time tomorrow; you can
remaining,both TCs and naifs do it now if and only do it for the first time tomorrow. Thus,
only if the currentperiod is one of the k best forgoing the reward today makes you delay
remaining periods given their current prefer- until period 3. When M = 2, the period-I so-
ences. Having more activities remaining, phisticate knows he will do the second activity
therefore, makes it more likely that they per- in period 2 if he does the first now, but he can
form an activity now. Part 2 of Proposition 6 force himself to do it in periods 2 and 3 if he
states that the qualitativebehavior of naifs rel- waits now.
ative to TCs in the multiactivity model is ex- Example 4 also illustrates that the simple
actly analogous to that in the one-activity comparison of Proposition 2-that for M =1
model. If costs are immediate, naifs procras- sophisticates always do it before naifs-does
tinate: They are always behind TCs in terms not extend to the multiactivity case. In Ex-
of activities completed so far. If rewards are ample 4 with M 2, sophisticates do it after
immediate, naifs preproperate:They are al- naifs. The intuition behind Proposition 2 was
ways ahead of TCs in terms of activities com- that sophisticates are correctly pessimistic
pleted so far. Hence, the present-bias effect about their utility from completing the activity
extends directly to the multiactivity setting; in the future, and are therefore less willing to
and again naifs exhibit the pure effects of wait than naifs. But for M > 1 the relevant
present-biasedpreferences. question is how pessimism affects the mar-
While the behavior of naifs in the multi- ginal utility of delaying one activity. As a re-
activity model is a straightforwardand intui- sult, there is no general result for the
tive analogue of their behavior in the implications of sophistication versus naivete.
one-activity model, the effects of sophistica- Example 4 shows for immediate rewards that
tion are significantly complicated. Consider sophistication can sometimes mitigate rather
the following example. than exacerbate preproperation.Likewise, for
immediate costs one can also find cases where
Example 4: Suppose rewards are immediate, sophistication exacerbates procrastination
T - 3, and - 1/2 for naifs and sophisticates. (and where sophisticates are worse off than
Let v= (6, 1, 21 ) and c= (0, 0, 0). naifs). These examples illustrate that, in gen-
If M 1, then T-, 1, T, = 2, and T,= 3. eral environments, identifying when sophisti-
If M = 2, then 0Q(2) = {2, 31, 0,,(2) - cation mitigates self-control problems and
{1, 2 ,and 0,(2) = {2, 3}. when it exacerbates them is more complicated
than in the one-activity model. It is still true
There are a couple of aspects of Example 4 that sophisticates are more pessimistic than
worth emphasizing. First, changing M dra- naifs about future behavior. But in more gen-
matically changes the behavior of sophisti- eral environments, comparisons of sophisti-
cates: While sophisticates always preproperate cates to naifs depends on whether pessimism
when there is one activity, they do not prepro- increases or decreases the marginal cost of
perate here with two activities. Hence, the an- current indulgence. As we discuss in Section
VII, in many contexts there are identifiable particulcareconomic applications, such as sav-
patterns as to how pessimism will affect in- ings and addiction.2
centives to behave oneself-but these patterns In ouirmodel, the behavior of naifs intui-
will not always correspond to the simple case tively and directly reflects their bias for the
of Proposition 2. present. We suspect this simplicity in predict-
We conclude this section by returningto a ing the effects of naive self-control problems
point made in Section III-that sophistication will hold in a broad arrayof economic models.
can lead a person to behave in ways that are Since consuming now yields immediate pay-
seemingly contrary to having present-biased offs whereas the increased future payoffs that
preferences. In Section III, we showed that so- saving allows is delayed, naifs will undersave
phisticates may do it before TCs even though in essentially any savings model; and since ad-
costs are immediate. In the following example, dictive activities involve yielding to some im-
sophisticates do things after TCs even though mediate desire today that has future costs,
rewards are immediate. naifs will overindulge in essentially any ad-
diction rnodel.
Example 5: Suppose rewards are immediate, In contrast to naifs, sophisticates in our
and ,B - 1/2 for naifs and sophisticates. Let model can behave in ways that seemingly con-
v = (12, 6, 11, 21) and c (0 0,0, 0). tradict having present-biased preferences. We
If M = 2, then 0,t(2) { 1, 4}, 0,,(2) saw in Section III that sophisticates may com-
{1, 3 }, and 0E(2) = 3, 4}. plete an unpleasant task before they would if
they had no self-control problem, and in Sec-
In Example 5, the situationbeginning in pe- tion VI that they may consume temptinggoods
riod 2 is identical to Example 4, and the in- later than they would if they had no self-
tuition for why sophisticates do it later than control problem. We suspect this complexity
TCs is related to the intuition of Example 4. in predicting the effects of sophisticated self-
The period-I sophisticate knows that if he has control problems will also hold more gener-
one activity left in period 2, he will do it in ally. Sophistication effects that operate in
period 2, while if he has two activities left in addition to, and often in contradiction to, the
period 2, he will wait until periods 3 and 4. present-bias effect can be quite significant. In
Hence, even though the period-I sophisticate's the realmlof saving, sophisticates can have a
most preferredperiods for doing it are periods negative marginalpropensity to consume over
1 and 4, he realizes he will not do it in period some ranges of income; and sophisticates can
4 if he does it in period 1. The choice for the sometimes save more than TCs (i.e., they can
period-I sophisticate is between doing it in pe- behave exactly opposite from what a present
riods 1 and 2 versus doing it in periods 3 and bias would suggest) 23 In the realm of
4. Of course situations like Example 5 are
somewhat special; but we do not feel they are
so pathological that they will never occur in
real-world environments (particularlyfor dis- 22'here has been a lot of previous research on time
crete choices). inconsistency in savings models; see, for instance, Strotz
( 1956), Phelps and Pollak ( 1968), Pollak ( 1968), Thaler
and Shefrin (1981), Shefrin and Thaler (1988, 1992),
VII. Discussionand Conclusion Laibson (1994, 1995, 1997), and Thaler (1994). Re-
cently, economists have proposed models of 'rational ad-
diction" (Becker and Murphy, 1988; Becker et al., 1991,
1994). These models insightfully formalize the essence of
Many economic applicationswhere present- (bad) addictive goods: Consuming more of the good today
biased preferences are clearly important can- decreases overall utility but increases marginal utility for
not readily be put into the framework of this consumption of the same good tomorrow. However, these
paper. Nonetheless, we feel our analysis pro- models a priori rule out the time-inconsistency and self-
vides some insight into such realms. In this control issues modeled in this paper, and which many ob-
servers consider importantin addiction.
section, we discuss some general lessons to 23 For simple examples of such behaviors, consider the
take away from our analysis, and illustrate following savings interpretationof a multiactivity model
how these general lessons might play out in with c - (0, 0, ..., 0): People have time-variant
lar is negligible. Then given wealth M E { $1, $2,..., $T} sophisticationmight affect the implications of self-control
you must decide in which periods to consume. With this problems. Strotz ( 1956) and Akerlof ( 1991 ) discuss how
savings interpretation,sophisticates have a negative mar- sophistication might help improve behavior. We suspect
ginal propensity to consume in Example 4: With wealth their discussion reflects the prevalent intuition that so-
$1, sophisticates consume $1 in period 1, while with phistication can only help, and in fact have found no ex-
wealth $2, sophisticates consume $0 in period 1. And so- plicit discussion anywhere of how awareness of
phisticates save more thanTCs in Example 5: With wealth self-control problems might hurt. That sophistication can
$2, TCs consume $1 in year 1 and save $1 (which is con- hurt you is, however, implicit in Pollak (1968). In the
sumed in year 4), while sophisticates consume $0 in year process of demonstrating a mathematical result, Pollak
1 and save $2 (which is consumed in years 3 and 4). shows that sophisticates and naifs behave the same for
Although Examples 4 and 5 use ratherspecial utility func- logarithmicutility. From this, it is straightforwardto show
tions, it is relatively straightforwardto find similar ex- that for utility functions more concave than the log utility
amples where utility functions are concave, increasing, function, sophisticates save more than naifs (i.e., sophis-
and differentiable. We suspect, but have not proven, that tication mitigates self-control problems), whereas for less
sophisticates will never save more thanTCs if utility func- concave utility functions, sophisticates save less thannaifs
tions are constant over time. (i.e., sophistication exacerbates self-control problems).
By analyzing the implications of present- result above, in either case U0( Tr,) 2( U"( r ),
biased preferences in a simple model, and pos- which implies U0(rt,) - U?(Tr,)? U0(rt,) -
iting some general lessons thatwill likely carry U0(T). Given the definition ofT, ss = N only
over to other contexts, we hope that our paper if fVTtc - CTt.K < pU0(T) or -((1 - f3)
will add to other researchin developing a trac- 'f 3)CTtC + U0(Ttr) < U(T ). Given the upper
table means for economists to investigate both bound on costs X, we must have U0(Tr,) -
the behavioral and welfare implications of U0 (Tr) < (( 1- f)1/f)X. It is straightforward
present-biasedpreferences. to show we can get arbitrarilyclose to this
bound, so sup(VC)[U? (Tr) -- U?(r,)] = ((1 -
f )f3)X . Hence, lim,61 (sup(v,c)[U0(Tt,) -
APPENDIX Uo(,r,) = 0.
(2) Fix 3 < 1. We will show that for any
PROOF OF PROPOSITION 1: E E (0, X) there exist reward/cost schedule
(1) We show that when costs are imme- combinationssuch that U?(rt,) - U0( Tn) =
diate, for any period if naifs do it then TCs 2X - s, from which the result follows. Choose
do it. Consider period t, and let t' y > 0 such thatf3 + y < 1. Let i be the integer
maxT,>t(vT - cT). Naifs do it in period t only satisfying (e)/(f3 + y)l < X - (?)/
ifIvt-ct 2- /(vt - ct'), orvt - (l/f6)ct 2 (/3 + y) + 1, and let] be the integer satisfying
vt,- c,; TCs do it in period t if vt - ct X - j((l - 3)/(f + y))X > 0 > X -
V- ct,; andvt - ct 2 vt - (1//6)ct for any (j + 1)((1 - )/(f + y))X. Consider the
3 ? 1. The result follows. following reward and cost schedules where
(2) We show that when rewards are im- T= i + j + 3 is finite:
mediate, for any period if TCs do it then naifs
do it. Consider period t, and let t' = v= (X,X, ... .,X,X- ((1- f)/(
maxT,>t(vT - cT). TCs do it in period t only if
vt - :t vt, - c,; naifs do it in period t if + y))X,X- 2((1 - /3)/(f3 + y))X,
-vt fct2 ? (vt, - ct), or (l/f3)vt - ct2
Vt' -ct,; and (l/f3)vt - ct ? vt - ctfor any . . ,X-j( - om/p + Y) )Y, O)
3 c 1. The result follows.
C= /:+ Y),) 1?( p + 7 )2 .. I
(1) We first argue that when costs are im- rewrite as ((1 - f)/f3)vT, + U0(,T) 2
mediate,for any t < t' suchthatss = s', = Y, U (Tt,). Given the upper bound on rewards
UO( t) 2 UO( t ). This follows because for any X, we have U0(Tt,) - U0(,T) ! ((1 - 3)/
t and t' min,>tIrIss = Y}, ss = Yonly if /3)X. Since the bound is easily achieved,
3Vt -ct 2 /(vt - ct), which implies vt - 5UP(VC)[JU (TtC) -
U0(,T)] =((1 -
be the integer satisfying ()/(3i) <X (?)/ For any /3and 6, choose E (F/3, 1), let
(
(/i + ' ), and let j be the integer satisfying X - v = (1,
k/(06), and c = (0, 0,
02/(/6262))
j((l - f3)/f)X> O X - (j + 1)((1 - 3)/ 0), and suppose rewards are immediate. Then
,1)X. Consider the following reward and cost naifs choose Tn 2 when facing v and c, but
schedules where T = i + j + 3 is finite: Tn =1 when facing v-' and c -T, and this vi-
olates independence of irrelevantalternatives.
V = (?, e/(f), I/(/ ). * I(f' ),X,
Loewenstein, George and Prelec, Drazen. nomic Inquiry, October 1988, 26(4), pp.
"Anomalies in IntertemporalChoice: Evi- 609-43.
dence and an Interpretation." Quarterly ___====. "Mental Accounting, Saving, and
Journal of Economics, May 1992, 107(2), Self-Control," in George Loewenstein and
pp. 573-97. Jon Elster, eds., Choice over time. New
O'Donoghue, Ted and Rabin, Matthew. "Doing York: Russell Sage Foundation, 1992, pp.
It Now or Later." Math Center Discussion 287-329.
Paper No. 1172, Northwestern University, Strotz, RobertH. "Myopia and Inconsistency
1996. in Dynamic Utility Maximization." Review
_____. "Incentives for Procrastinators." of Economic Studies, 1956, 23(3), pp.
Math Center Discussion Paper No. 1181, 165-80.
Northwestern University, 1997; Quarterly Thaler,RichardH. "Toward a Positive Theory
Journal of Economics, 1999 (forth- of Consumer Choice." Journal of Eco-
coming). nomic Behavior and Organization, March
Peleg, Bezalel and Yaari, Menahem E. "On the 1980, 1(1), pp. 39-60.
Existence of a Consistent Course of Action _ . "Some Empirical Evidence on Dy-
When Tastes Are Changing." Review of namic Inconsistency," in Richard H.
Economic Studies, July 1973, 40(3), pp. Thaler, ed., Quasi rational economics. New
391-401. York: Russell Sage Foundation, 1991, pp.
Phelps, E. S. and Poflak, Robert A. "On Second- 127-33.
Best National Saving and Game-Equilibrium . ''Psychology and Savings Policies."
Growth."Review of EconomicStudies, April American Economic Review, May 1994
1968, 35(2), pp. 185-99. (Papers and Proceedings), 84(2), pp.
Pollak,RobertA. "Consistent Planning." Re- 186-92.
view of Economic Studies, April 1968, Thaler, Richard H. and Loewenstein, George.
35(2), pp. 201-08. "Intertemporal Choice," in Richard H.
Prelec,Drazen."Decreasing Impatience: Defi- Thaler, ed., The winners curse: Paradoxes
nition and Consequences." Working Paper and anomalies of economic life. New York:
No. 90-015, HarvardUniversity, 1990. Free Press, 1992, pp. 92-106.
Schelling,ThomasC. "Egonomics, or the Art of Thaler,RichardH. and Shefrin,Hersh M. ""An
Self-Management." American Economic Economic Theory of Self-Control." Jour-
Review, May 1978 (Papers and Proceed- nal of Political Economy, April 1981,
ings), 68(2), pp. 290-94. 89(2), pp. 392-406.
. "Self-Command in Practice, in Pol- Watterson,Bill. The days are just packed: A
icy, and in a Theory of Rational Choice." Calvin and Hobbes collection. Kansas City,
American Economic Review, May 1984 MO: Andrews and McMeel, 1993.
(Papers and Proceedings), 74(2), pp. 1- Wertenbroch,Klaus. "Consumer Self-Control:
11. How Much Vice and Virtue Do We Buy?"
. "Self-Command: A New Disci- Mimeo, University of Chicago, 1993.
pline," in George Loewenstein and Jon Yaari, MenahemE. "Consistent Utilization of
Elster, eds., Choice over time. New York: an Exhaustible Resource, or, How to Eat an
Russell Sage Foundation, 1992, pp. 167- Appetite-Arousing Cake." Center for Re-
76. search in Mathematical Economics and
Shefrin, Hersh M. and Thaler, Richard H. "The Game Theory Research Memorandum No.
Behavioral Life-Cycle Hypothesis." Eco- 26, 1977.