The Journal of Socio-Economics 36 (2007) 36–47
The framing of ﬁnancial windfalls and
implications for public policy夽
Nicholas Epley ∗, Ayelet Gneezy
University of Chicago Graduate School of Business, 5807 South Woodlawn Avenue, Chicago, IL 60637, USA
Governments, employers, and companies provide ﬁnancial windfalls to individuals with some regularity.
Recent evidence suggests that the framing (or description) of these windfalls can dramatically inﬂuence
their consumption. In particular, income described as a positive departure from the status quo (e.g., as
a bonus) is more readily spent than objectively identical income described as a return to the status quo
(e.g., as a rebate). Such ﬁndings are consistent with psychological accounts of decision making and should
supplement existing economic models. These results have important implications for the marketing of such
windfalls, and discussion focuses particularly on implications for government tax policies.
© 2005 Elsevier Inc. All rights reserved.
JEL classiﬁcation: E210
Keywords: Framing; Rebate; Tax
In September 2001, the United States government returned a record 38 billion dollars to
its citizens in the form of tax rebates (Shapiro and Slemrod, 2003a), at least partly with the
stated goal of increasing consumers’ spending and stimulating the economy. As part of this
rebate, each tax paying American received a check for either $300, $500, or $600, depending on
his/her reported annual income. Although the American government regularly acts unilaterally,
it is certainly not alone in its occasional tendency to distribute ﬁnancial windfalls to individual
consumers. Governments around the globe distribute money to their constituents in various forms
of social services. Many employers dole out year-end “bonuses” to reward good performance.
夽This research was supported in part by NSF Grant #SES0241544. Portions of article, and some of the data described
herein, are adapted from Epley et al. (in press).
∗Corresponding author. Tel.: +1 773 834 1266.
E-mail address: email@example.com (N. Epley).
1053-5357/$ – see front matter © 2005 Elsevier Inc. All rights reserved.
N. Epley, A. Gneezy / The Journal of Socio-Economics 36 (2007) 36–47 37
Companies both big and small offer cash-back bonuses or rebates to entice sales. And some
people inherit money following the death of a relative or loved one who ran out of time before
they ran out of money. People certainly work hard for their regular paychecks, but most are
no stranger to unexpected and unearned ﬁnancial windfalls. Although ﬁnancial windfalls do not
comprise a signiﬁcant percentage of the average person’s total wealth, they are common targets for
interventions by outside agencies to alter a person’s consumption behavior. Understanding how
people respond to such windfalls is therefore of considerable interest for individual consumers
and public policy makers alike.
Economic models make reasonably strong predictions about how ﬁnancial windfalls should be
saved or spent. Some of these predictions are conﬁrmed in empirical tests, but others are not. Chief
among violations are cases in which the description or source of such income alters either the
likelihood that objectively identical income is saved or spent, or the kinds of items purchased with
the income. These violations of fungibility are predictable and difﬁcult to accommodate in rational
models of human behavior or decision making. This paper is intended to review new empirical
ﬁndings regarding the consumption of ﬁnancial windfalls, provide a psychological account of
how people respond to ﬁnancial windfalls that could supplement existing economic models, and
to discuss why these seemingly irrational responses might be of practical importance.
1.1. Economic models of spending and saving
No person would presumably desire to go from living like a prince one day to living like
a pauper the next, and traditional economic theories of spending and saving therefore assume
that people attempt to create the most consistent standard of living they can reasonably sustain
over their lifetime. This desire for consistent consumption is seen most prominently in the two
dominant economic theories of spending and saving—the Life-cycle (Ando and Modigliani, 1963)
and Permanent Income (Friedman, 1957) hypotheses.
The two theories propose that people attempt to maintain a constant standard of living across
ﬂuctuations in their income. To do so, people are assumed to save during periods of unusually high
income and dissave during periods of unusually low income. As a result, people should dissave
both early in life (when people are making less than their average lifetime salary) and late in life
(when people have fewer years to live), but should save money in mid-life when income is high
in order to fund future consumption.
Because of this preference for consistent consumption, temporary ﬂuctuations in income (such
as windfalls) should have little impact on consumption. In this respect, people should be no more
or less likely to spend windfalls, such as the 2001 U.S. Tax Rebate, than they would any other
kind of transient income. Windfalls, by their nature, represent a temporary boost in income and
therefore represent no permanent increase in one’s standard of wealth.
Both of these economic theories predict that individual decision makers will have long temporal
horizons, and will smooth consumption based on relatively long-term income levels. Empirical
evidence, however, suggests that people are more temporally myopic such that consumption is
overly sensitive to one’s current income. People are more likely to accept a risky gamble, for
example, after they have just won money in a gamble than after they have just lost money, a
ﬁnding known as the “house-money effect” (Thaler and Johnson, 1990). Across the lifespan,
the young and the old spend too much and the middle-aged spend too little given predictions
from the Life-cycle hypothesis (Thaler, 1980, 1992). Finally, people are more likely to spend
income when it is both unexpected and unearned, and is therefore less likely to be incorporated
into their overall wealth state. For example, participants in one particularly ingenious experiment
38 N. Epley, A. Gneezy / The Journal of Socio-Economics 36 (2007) 36–47
(Experiment 5, Arkes et al., 1994) either anticipated receiving $5 for their participation, or were
surprised to receive $5 once they arrived for the experiment. These participants were then sent off
to a college basketball game. Participants in the unexpected earnings condition spent more of their
earnings at the game than those in the expected income condition. These ﬁndings make it clear
that people treat income relatively piece-meal, considering each transaction more independently
than the actual fungibility of income would dictate.
This myopia in judgment and decision making has several interesting implications for how
people treat ﬁnancial windfalls. First, differing descriptions of objectively identical income should
systematically inﬂuence how people code the income and how they allocate it to different “men-
tal accounts” (Thaler, 1999). As we will show, simply altering the way a ﬁnancial windfall is
described can directly and dramatically inﬂuence spending and saving. Second, different sources
of ﬁnancial windfalls—say income inherited from one’s late grandmother versus income won at
the casino—may also inﬂuence spending and saving. In particular, these differing sources may
inﬂuence the kind of items purchased with money inherited from different sources (e.g., Levav
and McGraw, 2005). Although all dollars are created equal, one may feel a pang of reluctance at
spending grandma’s inheritance on a new sports car, but little reluctance spending casino earnings
doing the same.
In this paper we review new empirical ﬁndings related to the ﬁrst of these implications, and
present them not as a destructive critique of traditional economic models, but rather as a construc-
tive presentation of experimental results with important personal and public policy implications.
These data on ﬁnancial windfalls add to the growing body of literatures emphasizing the impor-
tance of incorporating basic psychological principles—judgmental myopia, in this case—into
economic theories of behavior.
2. Framing psychological windfalls
All income creates an objective gain in one’s absolute level of wealth. A $2000 tax return
means that a person is $2000 richer than he or she was the moment before. In order to detect
this change in income, of course, one needs to compare one’s current wealth state with the $2000
check to a prior state without the check and do some simple subtraction. This comparison process
is so obvious and rudimentary that it hardly seems worth mentioning, but paying close attention to
comparison processes has profound implications for how people code, and ultimately consume,
Traditional economic models of spending and saving do not highlight the importance of com-
parisons, but generally assume that behavior is guided by absolute assessments of wealth. Few
stimuli in the environment, however, can be evaluated absolutely but instead must be evaluated
in comparison to some standard or reference point. People are tall, stupid, or happy, for example,
only in comparison to others who are shorter, smarter, or sadder. As Kahneman and Tversky
(1979) put it, “our perceptual apparatus is attuned to the evaluation of changes or differences
rather than to the evaluation of absolute magnitudes.” Although income has an objective value
that may attenuate the importance of these comparison processes, there is little reason to assume
that comparisons with existing standards or reference points should have no inﬂuence at all.
In particular, note that ﬂuctuations in one’s wealth are usually detected by comparisons with
the status quo, or one’s current level of wealth, and changes in wealth are therefore perceived as a
relative gain or loss. A $2000 windfall means a person is relatively richer than a moment before,
and a $2000 robbery means a person is $2000 poorer than a moment before. But the immediate
N. Epley, A. Gneezy / The Journal of Socio-Economics 36 (2007) 36–47 39
status quo is not the only possible reference point. People have a lifetime of past wealth states
that could be used in the detection of change. The particular past wealth state used as a source of
comparison may therefore determine whether people experience a windfall as an objective gain
of some magnitude, or perhaps simply as a return to a previously better wealth state.
Notice that these comparisons suggest that income that creates an objective gain in wealth
may not always be perceived as a gain to the consumer. In particular, windfalls framed as a
gain from one’s current wealth state (e.g., as a bonus) may be perceived differently than income
framed as a returned to a previous wealth state (e.g., as a returned loss or rebate). Despite being
objectively identical, a “bonus” describes a positive change from the status quo whereas a “rebate”
subjectively describes a return to the status quo. If people evaluate income comparatively rather
than absolutely, they may feel like they have more income to spend—and therefore be more likely
to spend at least some of it—when it is described as a gain (e.g., as a bonus) than when it is
described as a returned loss (e.g., as a rebate).
To illustrate this point, imagine that you lose a $20 bill at the ofﬁce. In one case, a colleague
spots you the next day and gives you a $20 bill, noting that she saw it fall out of your wallet and
thought you would like it back. In another case, a colleague spots you the next day and gives you
$20, noting that she had just done well at the craps table and wanted to share some of her good
fortune with her favorite colleague. Although you are objectively—and unexpectedly—richer in
both cases, the ﬁrst case explicitly highlights one’s current lack of $20 compared to yesterday’s
possession of the same $20, whereas the latter case highlights a gain from a current wealth state.
Our strong suspicion is that people would be more likely to spend their $20 windfall in the latter
case when it is given than in the former case when it is returned. This suggests that windfalls may
be spent more readily when it is framed as a gain from one’s current wealth state (e.g., as a bonus)
than when it is framed as a return to a previous wealth state (e.g., as rebate or returned loss).
Epley et al. (in press) tested this framing hypothesis in a series of recent experiments. In each,
participants received unexpected income described as either a return to a previous wealth state
(e.g., as a “rebate”) or as a gain from a current wealth state (e.g., as a “bonus”). In one exper-
iment, for example, participants arrived in the laboratory and learned, quite unexpectedly, that
they were going to receive a $50 check as part of an experiment investigating how undergraduates
allocate ﬁnancial resources. All participants learned that this check was coming from a labora-
tory that—like most—was partially funded by students’ tuition dollars through the university’s
operating budget. Participants in the rebate condition were then told that “you are receiving this
tuition rebate because our lab has a surplus of funds,” that “we will contact you in 1 week to ask
you some questions about your tuition rebate,” and were asked if they had “any questions about
this tuition rebate.” All instructions were identical for participants in the bonus condition, except
that all three instances of “tuition rebate” were replaced with “bonus income.” Participants then
received their check and left—most wielding sizeable smiles. One week after this initial session,
participants were contacted by e-mail and asked to indicate how much of the $50 they had saved
and how much they had spent. No mention of “bonus” or “rebate” was used in this follow-up
As predicted, participants reported spending signiﬁcantly more of the $50 windfall when it
was described as a bonus (M= $22.04) than when it was described as a rebate (M=$9.55). In fact,
73% of participants in the rebate condition reported spending none of their $50 check, compared
to only 36% in the bonus condition. These results are signiﬁcant not only in statistical terms,
but in practical terms as well. Participants in the bonus condition reported spending, on average,
almost 2.5 times more of their objectively identical income compared with participants in the
40 N. Epley, A. Gneezy / The Journal of Socio-Economics 36 (2007) 36–47
The relatively simple design of this experiment is desirable because it does not restrict par-
ticipants’ spending in any way, nor were participants even informed that their spending would
ever be measured. It was designed, in fact, to be as similar as possible in an experimental context
to receiving a tax rebate check, such as from the federal government. It is not ideal, however,
because it relies exclusively on recalled behavior rather than on actual behavior. There was also
no instruction given to participants about what should be counted as spending and what should
be counted as saving, and it is possible that what participants chose to deﬁne as spending versus
saving somehow differed between the two conditions.
A second study retained some of the ecologically desirable aspects of this study with a poten-
tially more valid measure of behavior, by utilizing the same framing manipulation but asked
participants to document each expenditure of the windfall on a small accounting slip. At the end
of 1 week, participants were asked to return the accounting slip in the mail. Spending versus
saving was then coded by independent raters (unaware of the experimental hypotheses), instead
of by participants themselves. Despite these procedural changes, the results of this study were
conceptually identical to the ﬁrst. Participants in the bonus condition spent signiﬁcantly more
of their $50 windfall (M= $31.46) than participants in the rebate condition (M= $7.41). Overall,
75% of participants in the rebate condition saved all their $50 check compared to only 21% of
participants in the bonus condition.
This second experiment, however, went only part-way towards eliminating concerns regarding
self-reported behavior, and one ﬁnal experiment sought to rule out these concerns altogether by
creating a store in the laboratory from which participants could purchase a variety of items. As in
the preceding studies, participants received a ﬁnancial windfall—in this case $25—described as
either “bonus money” or “rebate money.” In contrast to previous experiments, however, partici-
pants were not given a check for the full amount but were instead told that they could choose to
spend any amount of the income on items available in a “lab store.” Participants were then shown
an array of 15 different items labeled with their sale prices, primarily consisting of university
memorabilia (e.g., mugs, pens, ID holders) and snack foods (e.g., soda, potato chips). Partici-
pants learned that these items were being sold at a 20% discount, that they could choose to spend
as much or as little of their $25 income as they wished, and that any unspent portion would be
given to them as a personal check. After studying the array, participants indicated how much of
their $25 they wished to spend, purchased their items, and received a personal check from the
experimenter for the unspent amount.
Consistent with the predictions and the results of the preceding experiments, participants spent
signiﬁcantly more of their income in the “lab store” when it was described as “bonus money”
(M= $11.16) than when it was described as “rebate money” (M= $2.43). Although participants
were generally more inclined to save their money than to spend it, this was especially true among
participants in the rebate condition—79% of participants in the rebate condition saved all of their
$25 income compared to only 16% of participants in the bonus condition.
The results of these three experiments suggest that decisions to spend or save ﬁnancial windfalls
may depend critically on the way those windfalls are described in comparison to one’s current
wealth. Income described as a gain from the status quo or a bonus is more likely to be spent than
income described as a return to a previous status quo or a rebate. These observed effects were not
small, but were substantial in all three of the experiments just described. The effect size (d)ofthe
basic framing manipulation was .62 in the ﬁrst experiment described, 1.22 in the second, and 1.28
in the last. This effect size is essentially the ratio of the mean difference between conditions and
the pooled standard deviation of the two conditions. An effect size over 1 therefore means that
the mean difference between conditions was larger than the pooled standard deviation of those
N. Epley, A. Gneezy / The Journal of Socio-Economics 36 (2007) 36–47 41
conditions, and the accepted categorization for a medium effect size is .5 and for a large effect size
is .8 (Cohen, 1988). Income framing in this context not only has a signiﬁcant effect on behavior,
it had a sizeable effect as well.
It is important to note that these experiments are relatively immune to many of the standard
concerns regarding the generalizability of psychological laboratory results to the broader pop-
ulation. These experiments involved real money rather than hypothetical scenarios or intuitive
judgments, measured spending in both fairly loosely controlled as well as more tightly controlled
conditions, used varying sizes of windfalls, and utilized different forms of payment—from per-
sonal checks to the equivalent of cash in the “lab store” experiment. They are not, of course,
immune to all concerns, and future experiments with a broader sample of participants, varying
amounts of payment, and alternative frames will undoubtedly identify important and interesting
moderators of windfall framing effects. What we know now is that describing a windfall as a
gain from a wealth state can increase spending compared to describing a windfall as a return to a
previous wealth state. What we don’t know is how, and how much, the size of this framing effect
is likely to vary from one moment to the next or from one participant to another. Such issues are of
critical importance in public policy decisions, and we look forward to the next round of research
to clarify these issues.
3. Accounting for income framing
Although the differences in spending and saving in the preceding experiments are clear, the
reasons for these differences are not. We have suggested here that a bonus describes a gain in
wealth whereas a rebate describes a returned loss and hence as no absolute gain in wealth. As
a result, people who are given a bonus feel like they have money to spend whereas those given
a rebate do not. A follow-up to the lab-study experiment tested this account directly by asking
participants to indicate the extent to which the windfall they received felt like additional income
versus returned income. In particular, participants were given a $25 windfall as described earlier,
given the opportunity to spend their income on items in the “lab store,” and then rated the extent
to which the windfall seemed like “extra money you received in additional whatever income you
would normally make this month” and the extent to which it seemed like “returned money that is
now being given back to you.”
As in the original study, participants in the bonus condition spent signiﬁcantly more of their
windfall (M= $7.63) than participants in the rebate condition (M= $1.63). More important, partic-
ipants in the bonus condition were also more likely to indicate that the windfall felt more like extra
money than returned money, and vice versa in the rebate condition. What is more, this difference
in the perception of wealth signiﬁcantly mediated the effect of windfall framing on spending.
This difference in perceived wealth, however, was only a partial mediator of the relationship
between windfall framing and spending, suggesting that additional mechanisms may be involved
as well. This follow-up experiment tested two additional mechanisms that appeared plausible,
but found support for neither. One alternative was that participants in the rebate condition were
less likely to spend income because a returned loss was perceived as more valuable than an
additional gain. If so, this may have made the rebate seem subjectively larger than the bonus, or
made the objects in the array seem relatively overpriced given then income’s value. Such a result
would be consistent with the asymmetrical gain/loss value function described by Prospect Theory
(Kahneman and Tversky, 1979), by which a returned loss should be seen as more valuable than
a simple gain. Such a result would also be consistent with the results of Gregory et al. (1993)
who found that participants stated—in a hypothetical scenario—that they would be willing to
42 N. Epley, A. Gneezy / The Journal of Socio-Economics 36 (2007) 36–47
pay more money for public policies framed as returned losses (e.g., restoring a wetland) than as
gains (e.g., creating a wetland). Participants’ responses in our studies, however, did not support
this prediction. There were no signiﬁcant differences in the extent to which the ﬁnancial windfall
“seemed like a large or a small amount,” in the extent to which the “objects seemed appropriately
priced,” or in the extent to which the “objects seemed like a good value.” We suspect this occurred
because the objective value of the income and the objective discount on the items for purchase
were so transparent that they allowed for little ambiguity in the subjective value of these goods.
This mechanism may therefore play a larger role in contexts where the objective value of an object
is more ambiguous (such as in those investigated by Gregory et al., 1993).
A ﬁnal alternative examined in this follow-up experiment was that returning money to par-
ticipants in the form of a rebate triggered thoughts about past spending and expenditures that a
bonus did not. People receive rebates, after all, only after they have spent some amount of money
in the ﬁrst place. Although the framing manipulation used in all of these experiments altered only
the label attached to the money itself and not its source or attention to prior expenses, describing
income as returned income may have led participants to think more carefully about the initial
expenditure, a thought that was not induced by the description of income as a bonus. In turn,
this may have inhibited spending by making participants feel like they have been spending too
much money recently, or have been spending their money unwisely and need to do more saving.
However, participants’ responses did not support this alternative either. There were no differences
in the amount participants reported spending this month compared to the average month, nor in
the extent to which participants reported that they “have been spending their money wisely.”
Failure to ﬁnd support for differences in the perceived value or sensitivity to past expenditures
does not mean that such mechanisms do not play a role in how people spend ﬁnancial windfalls,
and it is possible that different measures or different contexts may ﬁnd support for one or both
of these mechanisms. It is also likely that additional psychological factors, such as a person’s
mood (Lerner et al., 2004) or affect associated with a particular windfall (Levav and McGraw,
2005) signiﬁcantly inﬂuence the likelihood of spending psychological windfalls. However, the
set of experiments we just described did not investigate the potential role of these factors in the
framing of psychological windfalls. The most we can say for now is that participants who received
a windfall described as a bonus believed they had extra income that those who received income
described as a rebate did not.
4. Implications—narrow and broad
The framing effects on ﬁnancial windfalls that we have described join a growing body of
evidence demonstrating the importance of incorporating basic psychological principles into eco-
nomic behavior and decision making (e.g., Ariely et al., 2003; Camerer, 1999; Kahneman, 2003;
Thaler and Benartzi, 2004). One of psychology’s most basic insights is that the evaluation of
objects is based heavily on descriptions of objects rather than simply on objective features of
the objects themselves (Tversky and Koehler, 1994). Different descriptions of the same objective
events can therefore yield very different judgments that sometimes—as in the experiments just
described—contradict very basic assumptions of traditional economic models. In this particular
case, describing income as a gain from the status quo dramatically increased people’s propensity
to consume income relative to describing income as a returned loss.
All of the experiments we described involved cases in which a current ﬁnancial windfall could
be related to a past expense. We believe it is likely that relating a windfall to any kind of past
expense—be it time, effort, or non-monetary losses—would produce similar results as well. We
N. Epley, A. Gneezy / The Journal of Socio-Economics 36 (2007) 36–47 43
would expect, for instance, that a year-end windfall from an employer would more likely be saved
if it were described as a reward for the past year’s efforts than if it was simply described as a gift of
appreciation. Like the rebate described in relation to a past expense, income described in relation
to one’s past efforts may not feel like an absolute gain in the same way as income that appears
both unexpected and unearned, and may therefore not promote the same level of spending. This
may be part of the reason that unexpected and unearned ﬁnancial windfalls, per se, are more likely
to be spent than non-windfalls (Arkes et al., 1994).
What is more, the psychological mechanisms involved in the framing effects we have described
are likely to produce analogous results in other domains as well. Consider, for example, the
increased concern that would seem to arise from a 10% increase in the likelihood of a national
terrorist attack compared to a 10% return to last month’s level of risk. Or differences in the
likelihood of changing one’s diet after gaining ten pounds over the last month compared to
regaining ten pounds. Or the increased likelihood of “wasting” one’s time if a regularly scheduled
meeting is canceled adding an “extra” hour to one’s normal work day compared to a one-time
meeting being canceled that simply “returned” an hour to one’s work day. From calculating risk
to consuming time, the ability to frame events as a change from one’s current state versus a return
to a previous state seems common. Empirical extensions of the framing effects we described may
therefore be quite broad.
Perhaps the most important implications of windfall framing effects, however, are for govern-
ment tax policies. As mentioned earlier, one reasonably common economic strategy for stimulating
a local economy is to redistribute wealth to constituents. These government windfalls can take
many forms, from tax returns at the end of the ﬁscal year to formal tax cuts, but the most common
for stimulating the economy directly are tax rebates. Tax rebates are one-time windfalls distributed
according to a person’s overall wealth. The most recent of these tax rebates—and also the largest
the world has ever seen—was the U.S. tax rebate of 2001 in which 38 billion dollars were dis-
tributed to U.S. tax-payers in the form of $300, $500, or $600 checks. The economic logic of this
tax cut was fairly simple—spending is a function, at least in part, of a person’s absolute wealth,
so increasing the wealth of U.S. citizens will increase spending and stimulate economic growth.
According to the Bush administration’s council of economic advisors, this tax rebate “provided
valuable stimulus to economic activity in the short run” and “softened the recessionary headwinds
in 2001 that has helped to put the economy on the road to recovery in 2002” (cited in Shapiro and
Of course, when times are unusually bad, it’s useful to remember that times are likely to
get better by statistical chance alone, and it’s not entirely clear what data would conclusively
demonstrate a softening of recessionary headwinds. What is more, empirical data suggests that
the tax rebate might not have been quite as effective in stimulating short-term spending as the
Bush administration might have hoped. In one survey, for instance, only 22% of taxpayers reported
that they would spend their tax rebate check and the remaining vast majority reported that they
would save it (Shapiro and Slemrod, 2003b). Given that the average American has a difﬁcult time
saving even the smallest percentage of their regular income, these reported savings rates are quite
These reported savings rates are also reﬂected in macroeconomic data of actual behavior
(Shapiro and Slemrod, 2003a). The tax rebates of 2001 were distributed primarily during July,
August, and September. In the ﬁrst 6 months of 2001, the personal savings rate as percentage of
disposable personal income hovered around 2%, but nearly doubled over the following 3 months,
coinciding perfectly with the distribution of the rebate checks. Very similar ﬁndings were observed
following a very similar tax rebate in 1975, when savings rates spiked from approximately 10%
44 N. Epley, A. Gneezy / The Journal of Socio-Economics 36 (2007) 36–47
before the rebate to roughly 14% after the rebate. The results do not indicate that tax rebates did
not, in fact, soften the recessionary headwinds in 2001, but rather that tax rebates might not have
softened such headwinds as much as they could have.
Economic theory explains these increases in savings by noting that the rebates were not per-
manent and therefore should not have inﬂuenced consumers’ spending. This may well be true,
but the windfall framing effect suggest another possibility—that the tax rebates were coded as
a returned loss rather than as an additional gain, and hence did not stimulate spending for the
same reasons we saw in the rebate conditions described earlier. Some of the administration’s own
political rhetoric, in fact, seemed to encourage this kind of framing. When unveiling the proposed
rebate, for instance, President Bush argued that a budgetary surplus “should be returned to the
taxpayers who earned it” because “it’s the people’s money and government ought to be passing it
back after it’s met priorities” (Bush, 2001). Many clarion calls from politicians for decreases in
government spending and economic growth herald the need to “return the tax-payer’s money,” a
framing that might be a very ﬁne political strategy but a poor marketing strategy.
Scientists are obviously unable to manipulate the way governments distribute to its constituents,
butwe(Epley et al., in press) tried to do the next best thing by investigating how windfall framing
might inﬂuence people’s memory for spending of their 2001 tax rebate. Although decisions about
whether to spend or save income are superﬁcially distinct from one’s memory for spending and
saving, the reconstructive process of memory (Schacter et al., 1998) operates in much the same
way as the construction of preferences that precedes decision and choice (Slovic, 1995). Therefore,
the effect of framing on the construction of preferences for spending or saving should operate
similarly in the reconstructive process of memory.
To examine the role of framing on memory for spending the 2001 tax rebates, a sample of
Boston-area residents were recruited in public train stations several months after disbursement
of the rebates. All were ﬁrst asked whether they recalled receiving a check—$300, $500, or
$600—from the 2001 Tax Relief Act, and all did. Participants then read one of two descriptions
of the 2001 Tax Relief Act at the top of a questionnaire—one that described the checks as an
additional income resulting from a budget surplus that should be returned as bonus income,or
another that described it as tax surplus that should be returned as withheld income (i.e., as returned
income). In particular, those in the bonus condition read that “proponents of this tax cut argued
that the costs of running the government were lower than expected, resulting in a budget surplus”
that should be returned “as bonus income,” whereas participants in the returned income condition
read that “proponents of the this tax cut argued that the government collected more tax revenue
than was needed to cover its expenses, resulting in a tax surplus” that should be returned “as
withheld income.” The framing in the returned income condition, in fact, was paraphrased from
the Bush administration’s description of the rebate.
All participants were then asked to indicate the rebate amount their household received ($300,
$500, or $600), and what percentage of this money they recalled spending and what percentage they
recalled saving. As predicted, participants in the bonus condition recalled spending, on average, a
whopping 87% of their tax rebate whereas participants in the returned income condition recalled
spending, on average, only 25%. The similarity between this latter ﬁgure and the 22% predicted
spending ﬁgure reported by Shapiro and Slemrod (2003a) may be no coincidence, given the
similarity between the description in the returned income condition and the frame participants
were likely to hear of this tax rebate in their daily lives. This result is consistent with the possibility
that the low spending rates of tax rebates are, at least partly, a function of the way such rebates
are naturally described, and suggests that an alternate frame has at least some hope of inﬂuencing
N. Epley, A. Gneezy / The Journal of Socio-Economics 36 (2007) 36–47 45
The results of this study were replicated in a follow-up experiment conducted approximately 6
months later. This time the participants were 76 travelers in New York City’s Grand Central Station.
These participants were randomly assigned to the same conditions as in the previous experiment.
Once again, participants in the bonus condition recalled spending signiﬁcantly more of their tax
rebate (M= 76%) than those in the rebate condition (M= 41%). More important, participants in
this study were also asked the extent to which they perceived the rebate as a gain from one’s
current wealth—as “ ‘extra’ money that you received in addition to whatever you would normally
make this month”—and the extent to which they perceived it as a return to a previous wealth
state—as “money belonging to your original income that was temporarily withheld, and which
is now being given back to you.” Those in the bonus condition were more likely to perceive the
check as extra income and less likely to perceive it as returned income, compared to participants
in the rebate condition—a difference that signiﬁcantly mediated the relationship between windfall
framing and recalled spending.1
It is important to remember that these differences in recalled spending are clearly memory
errors and almost certainly do not reﬂect differences in actual spending, as participants were
randomly assigned to their respective conditions. Nevertheless, these results, in concert with the
behavioral experiments described earlier, suggest that altering the description of a tax rebate to
highlight the objective gain in income, may be an effective way to increase spending and stimulate
economic recovery. Altering the frame of such tax policies cost nothing, so at the very least any
inﬂuence on behavior would undoubtedly be an amazing return on the ﬁnancial investment.
Notice that similar logic can be applied to more permanent tax cuts as well. To the extent
that such permanent tax cuts are meant to increase consumers’ spending, care should be taken
to highlight the “bonus” income these tax cuts provide. Perhaps IRS checks in post-cut years
should include a report of what one’s taxes would have been without the new tax cut, and even
do the subtraction to keep the windfall salient in the decision maker’s mind and keep it from
simply melding into the background as part of one’s permanent income. We would predict, for
instance, that a tax return that included a check for $3000 plus a $500 bonus would be spent more
readily than a single check for $3500. Given that people evaluate individual transactions relatively
myopically, decoupling the “regular” tax return from a new “bonus” should facilitate spending of
the additional income. Indeed, decoupling expenses that are typically grouped together can have a
dramatic inﬂuence on consumption. For instance, people might not be wild about spending $500
on a hotel room, $125 on meals, and $75 on activities, but might be thrilled to spend $700 per day
on an all-expenses-paid hotel visit (Prelec and Loewenstein, 1998; Thaler, 1980; Van Boven and
1One secondary ﬁnding in both of these memory studies was that participants who received a smaller rebate check
($300) recalled spending a larger portion of it relative to participants who received a larger check ($500 or $600). This
difference was signiﬁcant in the ﬁrst memory-recall study we described and marginally signiﬁcant in the second. One dull
interpretation of this result is that those who received a larger check were wealthier, and therefore able to save more of
the check than those who received a smaller check and were less wealthy. This interpretation almost certainly has some
A more interesting interpretation is that people are simply more likely to spend small windfalls than larger windfalls.
This interpretation is consistent with theories of mental accounting, which suggest that small windfalls are not incorporated
into one’s overall income and are therefore not “booked” in the same way as larger windfalls (Thaler, 1999). Smaller
windfalls may be less likely to be deposited into one’s bank account, and instead be spent on smaller, frivolous purchases.
The propensity to consume a windfall may therefore be negatively correlated with its size relative to one’s permanent
income, exactly the result found in spending of windfalls given to Nazi Holocaust survivors by the German government
(Landsberger, 1966). Similar results were found for windfalls given to U.S. war veterans (Bodkin, 1966) and university
employees (Rucker, 1984). Whether governments could facilitate spending by simply distributing windfalls in smaller
amounts is therefore an intriguing but untested hypothesis.
46 N. Epley, A. Gneezy / The Journal of Socio-Economics 36 (2007) 36–47
Epley, 2003). None of these framing effects, of course, would alter ideological justiﬁcations for
various tax policies in the ﬁrst place, be they more liberal or more conservative, but they might
alter the impact of these policies on individual consumption.
As mentioned earlier, we believe these temporal framing effects are not restricted to tax policies
but apply to a variety of public policy domains as well. Most policy interventions aim to change
some current state, and many may therefore be described as creating a gain from the current status
quo versus returning to a previous status quo. Attempts to clean up the environment, for example,
could involve improving the current environment or returning to a cleaner environment of the
past. Diminishing terrorist threats could be described as reducing the high risk felt in 2005 or as
returning to the lower risks of 1995. And attempts to curb the obesity epidemic could be described
as a reduction in obesity rates of the current generation or as a return to the lower rates of previous
generations. In contrast to the studies we have described that attempt to change consumption,
many of these policies are attempts to change attitudes in order to improve the effectiveness of
these campaigns. As Gregory et al. (1993) demonstrated, returned losses in these domains can
be seen as more valuable than gains. Public policies that try to create positive attitudes for those
policies and increase compliance should therefore focus on returning to better days of the past
rather than on improving bad days of the present.
Although the direct implications of the work we have described on windfall framing for broader
public policy issues are somewhat speculative, we ﬁnd them worthy of consideration and further
empirical investigation. Returning speciﬁcally to ﬁnancial windfalls, surely government agencies
could conduct research of their own to investigate the impact of different descriptions of their
public policies on behavior in certain test markets before instituting one policy in particular. Such
market research is invaluable to the effectiveness of corporate campaigns, and would almost cer-
tainly be invaluable in public policy campaigns as well. Many politicians are experts at providing
just the right kind of “spin” on the available information, and in this respect might be well advised
to apply the same kind of framing strategies used to alter people’s attitudes into altering people’s
economic behavior. Given the amount of money spent on public tax policies alone, and the likely
impact these policies have on the overall economy and the population at large, time devoted to
understanding how decision makers code ﬁnancial windfalls would seem to be time well spent.
Ando, A., Modigliani, F., 1963. The life cycle hypothesis of saving: aggregate implications and tests. American Economic
Review 53, 55–84.
Ariely, D., Loewenstein, G., Prelec, D., 2003. Coherent arbitrariness: stable demand curves without stable preferences.
Quarterly Journal of Economics 118, 73–105.
Arkes, H.R., Joyner, C.A., Pezzo, M.V., Nash, J.G., Siegel-Jacobs, K., Stone, E., 1994. The psychology of windfall gains.
Organizational Behavior and Human Decision Processes 59, 331–347.
Bodkin, R.H., 1959. Windfall income and consumption. American Economic Review, 602–614.
Bush, G.W., 2001. Transcript of President Bush’s White House News Conference. New York Times, February 23, p. A10.
Camerer, C.F., 1999. Behavioral economics: reunifying psychology and economics. Proceedings of the National Academy
of Sciences 96, 10575–10577.
Cohen, J., 1988. Statistical Power Analysis for the Behavioral Sciences, second ed. Lawrence Erlbaum, New Jersey.
Epley, N., Mak, D., Idson, L., in press. Bonus or rebate? The impact of income framing on spending and saving. Journal
of Behavioral Decision Making.
Friedman, M., 1957. A Theory of the Consumption Function. Princeton University Press, Princeton, NJ.
Gregory, R., Lichtenstein, S., MacGregor, D., 1993. The role of past states in determining reference points for policy
decisions. Organizational Behavior and Human Decision Processes 55, 195–206.
Kahneman, D., 2003. A psychological perspective on economics. American Economic Review(Proceedings) 93, 162–168.
Kahneman, D., Tversky, A., 1979. Prospect theory: an analysis of decision under risk. Econometrica 47, 263–291.
N. Epley, A. Gneezy / The Journal of Socio-Economics 36 (2007) 36–47 47
Landsberger, M., 1966. Windfall income and consumption: comment. American Economic Review 56, 534–540.
Lerner, J.S., Small, D.A., Loewenstein, G., 2004. Heart strings and purse strings—carryover effects of emotions on
economic decisions. Psychological Science 15, 337–341.
Levav, J., McGraw, A.P., 2005. Emotional accounting: feelings about money and consumer choice. Unpublished
manuscript, Columbia University.
Prelec, D., Loewenstein, G., 1998. The red and the black: mental accounting of savings and debt. Marketing Science 17,
Rucker, M.H., 1984. Allocation of windfall income: a case study of a retroactive pay increase to university employees.
Journal of Consumer Affairs 18, 101–118.
Schacter, D.L., Norman, K.A., Koutstaal, W., 1998. The cognitive neuroscience of constructive memory. Annual Review
of Psychology 49, 289–318.
Shapiro, M., Slemrod, J., 2003a. Did the 2001 tax rebate stimulate spending? Evidence from taxpayers surveys. In:
Poterba, J. (Ed.), Tax Policy and the Economy. MIT Press, Cambridge, pp. 83–109.
Shapiro, M., Slemrod, J., 2003b. Consumer response to tax rebates. American Economic Review 93, 381–396.
Slovic, P., 1995. The construction of preference. American Psychologist 50, 364–371.
Thaler, R.H., 1980. Toward a positive theory of consumer choice. Journal of Economic Behavior and Organization 1,
Thaler, R.H., 1992. Savings, fungibility, and mental accounting. Journal of Economic Perspectives 4, 193–205.
Thaler, R.H., 1999. Mental accounting matters. Journal of Behavioral Decision Making 12, 183–206.
Thaler, R.H., Benartzi, S., 2004. Save more for tomorrow: using behavioral economics to increase employee savings.
Journal of Political Economy 112.1, S164–S187.
Thaler, R.H., Johnson, E.J., 1990. Gambling with the house money and trying to break even: the effects of prior outcomes
on risky choice. Management Science 36, 643–660.
Tversky, A., Koehler, D., 1994. Support theory: a nonextensional representation of subjective probability. Psychological
Review 101, 547–567.
Van Boven, L., Epley, N., 2003. The unpacking effect in evaluative judgments. Journal of Experimental Social Psychology